Definition and measurement

Oxidation rates of metals are measured in the way sketched in Figure 17.2: a thin sheet of the material (to give a lot of surface) is held at temperature T for an increasing time t, and the gain or loss in weight, Δm, is measured. If the oxide adheres to the material the sample gains weight in a way that is either linear (Δm ∝ t) or parabolic (Δ ∝ t^1/2) in time t; if instead the oxide is volatile, the sample loses weight linearly with time (Δm ∝ −t), as in Figure 17.3.

Figure 17.2 Measuring oxidation rates.

Figure 17.3 Oxidation rates: linear weight gain, parabolic gain and linear loss.

Polymers, too, oxidise, but in a more spectacular way. That brings us to flammability.

Flammability

Ceramics and glasses do not burn. Metals, if dispersed as a fine powder, are potentially combustible, but in bulk flammability is not normally an issue, even for magnesium. Most polymers, on the other hand, are inherently flammable, although to differing degrees: some burn spontaneously if ignited; others are self-extinguishing, burning only when directly exposed to flame.

There are several ways to characterise flammability. The most logical is the Limiting oxygen index (LOI): it is the oxygen concentration, in %, required to maintain steady burning. Fresh air has about 21% oxygen in it. A polymer with an oxygen index lower than 21% will burn freely in air; one with an oxygen index that is larger will extinguish itself unless a flame is played onto it—then it burns. Thus a high oxygen index means resistance to self-sustained burning. Less logical, but much used, is the Underwriters Laboratory (UL) rating in which a strip of polymer 1.6 mm thick is held horizontally (H) or vertically (V) and ignited. Its response is recorded as a code such as HB, meaning ‘horizontal burn’. The scales are compared in Table 17.1.

Table 17.1 Flammability ratings compared

Photo-degradation

You don’t have to set fire to a polymer for it to oxidise. Polymers and elastomers age when exposed to light (particularly UV) and oxygen, causing loss of strength, stiffness and toughness, discoloration and loss of gloss. This is countered by additives: antioxidants, light stabilisers and fluorescent whitening agents. Some of these are so universal that the ‘standard’ grade of the polymer already contains enough of one or another of them to give it acceptable resistance; PP, ABS, PS, PET, PMMA, PC, nylons and PU all need UV protection.

Mechanisms of oxidation

The rate of oxidation of most metals in dry air at room temperature is too slow to be an engineering problem; indeed, slight oxidation can be beneficial in protecting metals from corrosion of other sorts. But heat them up and the rate of oxidation increases, bringing problems. The driving force for a metal to oxidise is its free energy of oxidation—the energy released when it reacts with oxygen, but a big driving force does not necessarily mean rapid oxidation. The rate of oxidation is determined by the kinetics (rate) of the oxidation reaction, and that has to do with the nature of the oxide. When any metal (with the exception of gold, platinum and a few others that are even more expensive) is exposed to air, an ultra-thin surface film of oxide forms on it immediately, following the oxidation reaction

(17.1)

The film now coats the surface, separating the metal beneath from the oxygen. If the reaction is to go farther, oxygen must get through the film.

The weight-gain shown in Figure 17.3 reveals two different types of behavior. For some metals the weight gain per unit area, Δm, is linear, and this implies that the oxidation is progressing at a constant rate:

(17.2)

where k_ℓ (kg/m².s) is the linear kinetic constant. This is because the oxide film is porous or cracks (and, when thick, spalls off) and does not protect the underlying metal. Some metals behave better than this. The film that develops on their surfaces is compact, coherent and strongly bonded to the metal. For these the weight gain per unit area of surface is parabolic, slowing up with time, and this implies an oxidation rate with the form

(17.3)

where k_p (kg²/m⁴.s) is the parabolic kinetic constant.

In these equations Δm is the weight gain per unit area—that means the weight of the oxygen that has been incorporated into the oxide. The mass of the oxide itself is greater than this because it contains metal atoms (M) as well as oxygen (O). If the formula of the oxide is MO, the mass per unit area of the oxide is

where A_m is the atomic weight of the metal and A_o is that of oxygen. The thickness of the oxide is simply x = m_ox/ρ_ox where ρ_ox is the oxide density. Using equation (17.3) we find

(17.4)

The oxide film, once formed, separates the metal from the oxygen. To react farther either oxygen atoms must diffuse inward through the film to reach the metal or metal atoms must diffuse outward through the film to reach the oxygen. The driving force is the free energy of oxidation, but the rate of oxidation is limited by the rate of diffusion, and the thicker the film, the longer this takes.

Figure 17.4 shows a growing oxide film. The reaction creating the oxide MO

Figure 17.4 Oxidation mechanisms (a) growth by metal diffusion and electron conduction, (b) growth by diffusion of oxygen and holes.

goes in two steps. The metal first forms an ion, M²⁺ say, releasing electrons:

The electrons are then absorbed by oxygen to give an oxygen ion:

The problem is that the first of these reactions occurs at the metal side of the oxide film whereas the oxygen reaction is on the other side. Either the metal ions and the electrons must diffuse out to meet the oxygen, or the oxygen and electron holes (described in Chapter 14) must diffuse in to find the metal. If the film is an electrical insulator, as many oxides are, electrons cannot move through it, so it is the oxygen that must diffuse in. The concentration gradient of oxygen is that in the gas C_o divided by the film thickness, x. The rate of growth of the film is proportional to the flux of atoms diffusing through the film, giving

(17.5)

where D is the diffusion coefficient, D_o and Q_d are the pre-exponential and the activation energy for oxygen diffusion. Integrating gives

This has the same form as equation (17.4) with

(17.6)

This explains why oxidation rates rise steeply with rising temperature, and why the growth is parabolic. The most protective films are those with low diffusion coefficients, and this means that they have to have high melting points. This is why the Al₂O₃ oxide film on aluminum, the Cr₂O₃ film on stainless steel and chrome plate and the SiO₂ film on high silicon cast iron are so protective.

Not all oxides grow with parabolic kinetics, as we have seen. Those that show linear weight gain do so because the oxide that forms is not compact, but has cracks or spalls off because of excessive volume change, leaving the fresh surface continually exposed to oxygen. Linear weight loss, the other behaviour, occurs when the oxide is volatile, and simply evaporates as it forms.

Flammability: how do polymers burn and how do you stop them?

The combustion of a polymer is an exothermic reaction in which hydrocarbons are oxidised to CO₂ and H₂O. That sounds simple, but it isn’t. Combustion is a gas phase reaction; the polymer or its decomposition-products must become gaseous for a fire to begin. When you light a candle you are melting the wax and raising it to the temperature at which it pyrolyses (400–800 °C) forming gaseous hydrocarbon decomposition products. These gasses react in the flame to produce heat, which melts and pyrolyses more wax, keeping the reaction going.

A fully developed fire results when an ignition source like a spark or cigarette ignites combustible material such as paper. Its heat radiates out causing other combustible materials (particularly polymers and fabrics) to decompose into a flammable gas mix. Flashover (Figure 17.5) occurs when these gasses ignite, instantly spreading the fire over the entire area and producing temperatures of over 1000 °C.

Figure 17.5 The temperature rise during a fire.

Combustion involves the reaction of free-radicals. At high temperatures hydrogen, together with one electron from its covalent bond (symbol H^·) is freed from a carbon atom of the polymer molecule, leaving a highly reactive free radical R^· in the gas. The hydrogen radical reacts with oxygen and the hydrocarbon radical to give CO₂ and heat, releasing the H^· again to propagate the reaction further.

Flame retardants work in one of two ways. Some scavenge the free radicals, tying them up harmlessly and so retarding or snuffing out the combustion reaction. These are usually compounds containing chlorine or bromine; they, too, pyrolyse to give free radicals Cl^· and Br^· that attach themselves to the hydrocarbon radical, removing it from the reaction. Others work by creating a protective layer of water vapor between the solid polymer and the gaseous decomposition products, limiting heat transfer, cooling it and reducing pyrolisation. Typical among these is the addition of Mg(OH)₂ that decomposes at about 300 °C, releasing H₂O and leaving inert MgO. Polymers containing flame retardants are identified by adding ‘FR’ to their names.

What causes photo-degradation?

As we have seen, commercial polymers are long chain, high molecular weight hydrocarbons. When exposed to radiation chemical reactions are triggered that change their chemical composition and molecular weight. These reactions, called photo-oxidation or photo-degradation, also create free radicals. They trigger a change in the physical and optical properties of the polymer, making them brittle and, if transparent, turning them white or gray. Once this starts it sets off a chain reaction that accelerates degradation unless stabilisers are used to interrupt the oxidation cycle. Heat, too, can trigger degradation in an oxygen-containing atmosphere.

Ultraviolet absorbers such as benzophenone or the benzotriazole work by preferentially absorbing UV radiation, following the grandly named Beer–Lambert law, which states the obvious: that the amount of UV radiation absorbed increases with increases in the sample thickness and stabiliser concentration. In practice, high concentrations of absorbers and sufficient thickness of the polymer are required before enough absorption takes place to effectively retard photo-degradation. Hindered amine stabilisers (HALS) are a more efficient way to stabilise against light-induced degradation of most polymers. They do not absorb UV radiation, but act to inhibit degradation of the polymer; their efficiency and longevity are due to a cyclic process wherein the HALS are regenerated rather than consumed. Low concentrations are enough to give good stabilisation.

Aqueous solutions, acids and alkalis

Even when pure, water causes corrosion if oxygen is available. The rusting of steel in fresh water, alone, presents enormous economic and reliability problems. Add dissolved ions (like NaCl) and the corrosion rate increases dramatically.

Strong acids like H₂SO₄ (sulfuric acid), HNO₃ (nitric acid) and HCl (hydrochloric acid) are widely used in the chemical industry. So too are strong alkalis like NaOH (caustic soda). Household products contain some of these in dilute form, but in the home it is organic acids like vinegar (acetic acid) and milder alkalis like washing soda (sodium carbonate) that are more commonly found.

Dunk a metal into an acid and you can expect trouble. The sulfates, nitrates, chlorides and acetates of most metals are more stable than the metal itself. The question is not whether the metal will be attacked, but how fast. As with oxidation, this depends on the nature of the corrosion products and on any surface coating the metal may have. The oxide film on aluminum, titanium and chromium protects them from some acids provided the oxide itself is not damaged. Alkalis, too, attack some metals. Zinc, tin, lead and aluminum react with NaOH, for instance, to give zincates, stannates, plumbates and aluminates (as anyone who has heated washing soda in an aluminum pan will have discovered).

Ways of combating corrosion by water, acids and alkalis are described in Section 17.7. Corrosion, however, can sometimes be beautiful: the turquoise patina of copper on roofs and the rich browns of bronze statues are created by corrosion.

Oils and organic solvents

Organic liquids are ubiquitous. We depend on them as fuels and lubricants, for cooking, for removing stains, as face-cream, as nail varnish and much more—any material in service will encounter them. Metals, ceramics and glasses are largely immune to them, and some polymers can tolerate some organic liquids without problems. But not all. The identification of solvents that are incompatible with a given polymer is based largely on experience, trial and error and intuition guided by such rules as ‘like dissolves like’, meaning that nonpolar solvents like benzene (C₆H₆) or carbon tetrachloride (CCl₄) tend to attack nonpolar polymers like polypropylene or polyethylene, while polar solvents such as acetone (CH₃COCH₃) tend to attack polar polymers like PMMA and nylon.

Ions in solution and pH

Pure water, H₂O, dissociates a little to give a hydrogen ion, H⁺, and a hydroxyl ion, OH⁻:

(17.7)

The product of the concentrations of the two ions is constant: increase one, and the other falls. This is known as Law of Mass Action:

(17.8)

where the square brackets mean ‘molar concentration’. In pure water there are equal numbers of the two types of ion, [H⁺] = [OH⁻], and the value of the constant, when measured, is 10⁻¹⁴. Thus the molar concentration of both ion-types is 10⁻⁷; one H₂O molecule in 10⁷ is ionised. The pH of the ionised water is defined as the negative of the log of the hydrogen ion concentration:

(17.9)

so, for pure water, pH = 7.

Acids dissociate in water to give H⁺ ions. Sulfuric acid, for instance, dissociates

(17.10)

This pushes up the concentration [H⁺] and, because of equation (17.9), it pulls down the concentration [OH⁻]; weak acids have a pH of 4–6; strong ones a pH up to 0. Alkalis do the opposite. Sodium hydroxide, for example, dissociates when dissolved in water, to give OH⁻ ions:

(17.11)

Weak alkalis have a pH of 8–10, strong ones a pH up to 13 (Figure 17.6).

Figure 17.6 The scales of pH and pOH.

Corrosion by acids and alkalis is an electro-chemical reaction. One-half of this is the dissociation reaction of a metal M into a metal ion, M^z+, releasing electrons e⁻

where z, an integer of 1, 2 or 3, is the valence of the metal. Acidic environments, with high [H⁺] (and thus low pH) stimulate this reaction; thus a metal such as copper, in sulfuric acid solution, reacts rapidly:

(17.12)

Some metals are resistant to attack by some acids because the reaction product, here CuSO₄, forms a protective surface layer; thus lead-lined containers are used to process sulfuric acid because lead sulfate is protective.

Alkalis, too, cause corrosion via an electro-chemical reaction. Zinc, for instance, is attacked by caustic soda via the steps

Fresh and impure water

Corrosion, as we have just seen, is the degradation of a metal by an electro-chemical reaction with its environment. Figure 17.7 illustrates in more detail the idea of an electro-chemical reaction. If a metal is placed in a conducting solution like salt water, it dissociates into ions, releasing electrons, as the iron is shown doing in the figure, via the anodic reaction

Figure 17.7 Ionisation.

(17.13)

The electrons accumulate on the iron giving it a negative charge that grows until the electrostatic attraction starts to pull the Fe⁺⁺ ions back onto the metal surface, stifling further dissociation. At this point the iron has a potential (relative to a standard, the hydrogen standard) of −0.44 volts. Each metal has its own characteristic potential (called the standard reduction potential), as plotted in Figure 17.8. The extra electrons enter the band structure of the metal just above the Fermi level, so the energy associated with the Fermi level determines how strongly they are held.

Figure 17.8 Standard reduction potentials of metals.

If two metals are connected together in a cell, like the iron and copper samples in Figure 17.9, a potential difference equal to their separation on Figure 17.8 appears between them. The corrosion potential of iron, −0.44, differs from that of copper, +0.34, by 0.78 volts, so if no current flows in the connection the voltmeter will register this difference. If a current is now allowed, electrons flow from the iron (the anode) to the copper (the cathode); the iron ionises (that is, it corrodes), following the anodic reaction of equation (17.13) and—if the solution were one containing copper sulfate—copper ions, Cu⁺⁺, plate out onto the copper following the cathodic reaction

Figure 17.9 A bi-metal corrosion cell. The corrosion potential is the potential to which the metal falls relative to a hydrogen standard.

(17.14)

Suppose now that the liquid is not a copper sulfate solution, but just water (Figure 17.10). Water dissolves oxygen, so unless it is specially de-gassed and protected from air, there is oxygen in solution. The iron and the copper still dissociate until their corrosion potential-difference is established but now, if the current is allowed to flow, there is no reservoir of copper ions to plate out. The iron still corrodes but the cathodic reaction has changed; it is now the hydrolysis reaction

Figure 17.10 A bit-metal cell containing pure water in which oxygen can dissolve.

(17.15)

While oxygen can reach the copper, the corrosion reaction continues, creating Fe⁺⁺ ions at the anode and OH⁻ ions at the cathode. They react to form insoluble Fe(OH)₂, which ultimately oxidises further to Fe₂O₃.H₂O, rust.

Thus connecting dissimilar metals in either pure water or water with dissolved salts is a bad thing to do: corrosion cells appear that eat up the metal with the lower (more negative) corrosion potential. Worse news is to come: it is not necessary to have two metals: both anodic and cathodic reactions can take place on the same surface. Figure 17.11 shows how this happens. Here an iron sample is immersed in water with access to air. The part of the sample nearest the water surface has an easy supply of oxygen; that farther away does not. On the remoter part the ionisation reaction of equation (17.13) takes place, corroding the iron and releasing electrons that flow up the sample to the near-surface part where oxygen is plentiful, where they enable the hydrolysis reaction of equation (17.15). The hydrolysis reaction has a corrosion potential of +0.81 volts—it is shown in Figure 17.8—and the difference between this and that of iron, −0.44 volts, drives the corrosion. If the sample could be cut in two along the broken line in Figure 17.11 and a tiny voltmeter inserted, it would register the difference: 1.23 volts.

Figure 17.11 A corrosion cell created by differential access to oxygen.

This differential oxidation corrosion is one of the most usual and most difficult to prevent: where there is water and a region with access to oxygen and one that is starved of it, a cell is set up. Only metals above the hydrolysis reaction potential of +0.81 volts in Figure 17.8 are immune.

Selective corrosion

Often, wet corrosion, instead of being uniform, occurs selectively, and when it does, it can lead to failure more rapidly than the uniform rate would suggest. Stress and corrosion acting together (‘stress-corrosion’ and ‘corrosion-fatigue’) frequently lead to localised attack, as do local changes in microstructure such as those at a welded joint. Briefly, the localised mechanisms are these.

Intergranular corrosion occurs because grain boundaries have chemical properties that differ from those of the grain, because the disregistry there gives these atoms higher energy. Pitting corrosion is a preferential attack that can occur at breaks in the natural oxide film on metals, or at precipitated compounds in certain alloys. Galvanic attack at the microstructural level appears in alloys with a two-phase microstructure, made up of regions alternately of one microstructure and of another. The two regions will, in general, lie at slightly different points on the reduction potential scale. If immersed in a conducting solution, thousands of tiny corrosion cells appear, causing the phase that lies lower in reduction potential to be eaten away.

Stress corrosion cracking is accelerated corrosion, localised at cracks in loaded components. The stress breaks the protective oxide film and the elastic energy at and near the crack tip stimulates attack there. The result is that cracks grow under a stress intensity K_scc the is far below K_1c. Examples are brass in ammonia, mild steel in caustic soda and some aluminum and titanium alloys in salt water. Corrosion fatigue refers to the accelerated rate at which fatigue cracks grow in a corrosive environment. The endurance limit of some steels in salt water is reduced by a factor as large as 4. The rate of fatigue crack growth, too, increases to a level that is larger than the sum of the rates of corrosion and of fatigue acting together.

Design: material choice

Materials differ greatly in their vulnerability to attack by corrosive media, some surviving unscathed whereas others, in the same environment, are severely attacked. There are no simple rules or indices for predicting susceptibility. The best that can be done is a ranking, based on experience. Table 17.3 lists the vulnerability of common engineering materials to 23 commonly encountered fluids, on a four-point scale from A (excellent resistance to attack) to D (unacceptably rapid attack). Tables like these (or their computer-based equivalent¹) allow screening to eliminate materials that might present severe problems. But, as explained in the introduction to this chapter, there is more to it than that. The best choice of material depends not just on the environment in which it must operate, but—for economic reasons—on the application: it may be cheaper to live with some corrosion (providing regular replacement) than to use expensive materials with better intrinsic resistance. And there is the potential for protection of vulnerable materials by coatings, by inhibitors or by electro-chemical means. Experience in balancing these influences has led to a set of ‘preferred’ materials for use in any given environment. Table 17.4 presents these for the 23 environments that appear in the tables of this chapter.

Table 17.3 Corrosion-resistance ranking of materials in common environments

Table 17.4 Preferred materials and coatings for given environments

Design: geometry and configuration

Here are the Dos and Don’ts:

• Allow for uniform attack. Nothing lasts forever. Some corrosion can be tolerated, provided it is uniform, simply by allowing in the initial design for the loss of section over life.
• Avoid fluid trapping (Figure 17.13). Simple design changes permit drainage and minimise the trapping of water and potentially corrosive dirt.
• Suppress galvanic attack (Figure 17.14). Two different metals, connected electrically and immersed in water with almost anything dissolved in it, create a corrosion cell like that of Figure 17.9 or 17.10. The metal that lies lower on the scale of reduction potential is the one that is attacked. The rate of attack can be large and localised at or near the contact. Attaching an aluminum body shell to a steel auto chassis risks the same fate, as incautious car-makers have found. Even the weld metal of a weld can differ from the parent plate in reduction potential enough to set up a cell. The answer is to avoid bi-metal couples if water is around or, when this is impossible, to insulate them electrically from each other and reduce the exposed area of the nobler metal to minimise the rate of the cathodic reaction.
• Avoid crevices (Figure 17.15). Crevices trap moisture and create the conditions for differential-aeration attack. Figure 17.15 shows riveted lap joints (the rivet, of course, made from the same metal as the plate). If water can get under the plate edge, it will. The water-air surface has free access to oxygen, but the metal between the plates does not. The result is corrosion of the joint surface, just where it is least wanted. The answer is to join by welding or soldering, or to put sealant in the joint before riveting. This is not just a joint problem—differential aeration is very hard to avoid. It is the reason that the legs of jetties and piers rust just below the water line, and why some edges of tanks are more severely rusted than others.
• Consider cathodic protection (Figure 17.16). If attaching a metal to one lower in reduction potential causes the lower one to corrode, it follows that the upper one is protected. So connecting buried steel pipe-work to zinc plates sets up a cell that eats the zinc but spares the pipes. The expensive bronze propellers of large ships carry no surface protection (the conditions are too violent for it to stay in place). Copper and bronze lie above steel in the reduction-potential table and electrical isolation is impossible. Bronze connected to steel in salt water is a recipe for disaster. The solution is to shift the disaster elsewhere by attaching zinc plates to the hull around the propeller shaft. The zinc, in both examples, acts as a sacrificial anode protecting both the bronze and the steel. The term ‘sacrificial’ is accurate—the zinc is consumed—so the protection lasts only as long as the zinc, requiring that it be replaced regularly.
• Beware of stress corrosion and corrosion-fatigue. Stresses arising from design loads can be calculated; residual stress from cold work or from thermal contraction in the heat-affected zone of welds are not so easy to predict. Their interaction with a corrosive environment can be dangerous, giving localised stress-corrosion cracking, or adding to cyclic loads to give corrosion-fatigue. Internal stresses cannot exceed the yield strength of the material (if they did the material would yield) so, as a rule of thumb, select materials that have as low a yield strength (or hardness) consistent with the other constraints of the design.
• Design for inspection and maintenance. If you can’t reach it, you can’t fix it. Design for durability ensures that parts liable to corrosion are easy to inspect, clean and, if necessary, replace.

Figure 17.13 Design changes to prevent trapped fluids reduce the rates of corrosion.

Figure 17.14 Design changes to avoid or minimise galvanic attack.

Figure 17.15 Design changes to avoid crevice corrosion.

Figure 17.16 (a) Protection of steel pipes by a zinc sacrificial anode. (b) Galvanised steel plate; the zinc protects the steel even when scratched.

Coatings

Metals, as we have said, are reactive. Only gold and a couple of other precious metals are stable as metals rather than as oxides. Coatings allow reactive metals—and particularly steels, the backbone of mechanical engineering—to be used in environments in which, uncoated, they would corrode rapidly. They work in more than one way.

• Passive coatings separate the material from the corrosive environment. Some, like chrome, nickel or gold plate, are inherently corrosion resistant. Polymeric coatings—paints and powder-coats—provide an electrically insulating skin that stifles electro-chemical reactions by blocking the passage of electrons. Ceramic and glass coatings rely on the extreme chemical stability of oxides, carbides and nitrides to encase the metal in a refractory shell. Passive coatings work only if they are perfect (and perfection, in this world, is unusual); if scratched, cracked or detached, corrosion starts.
• Active coatings work even when damaged. Some, like zinc, are metals that lie low on the reduction-potential scale of Figure 17.8. Zinc on steel acts as a sacrificial coat, becoming the anode of the corrosion cell, leaving the underlying steel unscathed (Figure 17.16(b)). Polymeric coatings can be made to work in this way by dispersing zinc powder in them. Others contain corrosion inhibitors (described later) that leach into the corrosive medium, weakening its potency. Cement coatings inhibit attack by providing an alkaline environment.
• Self-generated coatings rely on alloying, almost always with chromium, aluminum or silicon, in sufficient concentrations that a film of Cr₂O₃, Al₂O₃ or SiO₂ forms spontaneously on the metal surface. Stainless steels, aluminum bronzes and silicon iron rely on protection of this sort. When scratched, the film immediately regrows, providing ongoing protection.

With this background, glance at Table 17.5. It lists coatings all of which you will have encountered even if you didn’t recognise them at the time. Modern engineering is totally dependent on them. They are applied in a number of ways, some illustrated in Figure 17.17.

• Metallic coatings rely on the intrinsic stability of the coat (gold); on its ability to form a self-generated, protective, oxide (chromium, aluminum, chromides and aluminides); or on its ability to provide cathodic protection (zinc, cadmium). Galvanized iron is steel sheet with a thin coating of zinc. The zinc acts as a sacrificial anode: if scratched, the exposed steel does not rust because the zinc protects it even when it does not cover it. Other platings look good but do not protect as well. A plating of copper sets up a cell that works in the opposite direction, eating the iron at a scratch rather than the copper. Chromium plating looks as if it should protect since Cr is below Fe in the table, but the chromium protects itself so well with its oxide film that it becomes passive and no cell is set up.
• Polymeric coatings rely either on a solvent that, after application, evaporates (acrylic, alkyds) or on a polymerisation reaction triggered by oxygen (linseed-oil based paints) or by an added catalyst (epoxies). Painting with solvent-based paints was, until recently, the most widely used method, but the evaporating solvent is generally toxic; it is a VOC (a volatile organic compound) and, for health reasons, their use is discouraged. Water-based paints overcome the problem but do not yet produce quite as good a paint film. Polymer powder coating works by flame-spraying the polymer or dipping the component when hot into a fluidised bed of polymer powder causing a thick film of polymer to melt onto the surface. While intact, paint films and polymer powder coats work well, but as soon as the coating is damaged the protection ceases unless the underlying metal has a subcoat of something (like zinc) that protects it.
• Ceramic and glass (enamel) coatings have to be applied hot, sufficiently so to sinter the particles of ceramic or to melt the glass onto the surface of the metal. Only cement coatings are applied cold.

Table 17.5 Coatings to enhance durability

Figure 17.17 Ways of applying coatings. (a) Paint spraying. (b) Hot dip galvanising. (c) Electroplating. (d) Metal flame spraying. (e) Polymer powder spraying. (f) Enamelling.

Corrosion inhibitors

Corrosion inhibitors are chemicals that, when dissolved or dispersed in a corrosive medium, reduce (‘inhibit’) the rate of attack. Some work (like indigestion tablets) by lowering the pH or by coating the part with a gooey film. Inhibitors suppress either the anodic or the cathodic step of the corrosion reaction (see Section 17.6). The choice of inhibitor depends both on the environment and on the material to be protected. Table 17.6 lists the more common ones and the metals and environments for which they work best. Many have alarmingly polysyllabic names that only specialised corrosion consultants try to remember. The things that are useful to know are itemised here.

• PTFE ((CF₂ − CF₂)_n) is used as a wide-spectrum inhibitor in hydrocarbon fuels, solvents and lubricants, most commonly in engine oil. It works by forming a protective layer over component surfaces, shielding them from full contact with the medium. Corrosion rates are reduced by up to 90%. Inadequate additions of PTFE lead to incomplete coverage and pitting corrosion, worsening the situation and leading to premature failure.
• Calcium bicarbonate (Ca(HCO₃)₂) is used in central heating, tap water and water cooling systems, where it lowers the pH and thus inhibits acidic attack. It makes the water hard, causing calcium deposits (lime scale) when water boils.
• Thiourea (C(NH₂)₂S) is an organic compound with a composition like that of urea but with the oxygen atom replaced by sulfur. It is the basic component of a range of organic anodic inhibitors that work by raising the activation energy of the corrosion reaction, thus decreasing the corrosion rates. Small additions can reduce the corrosion rate in hydrochloric acid by as much as 97%. Thiourea, however, is potentially carcinogenic, limiting its use to systems that do not involve human contact.
• Arsenic (III) oxide (As₂O₃) inhibits acidic attack of metals and, because it is an anodic inhibitor, it works well in a wide variety of acids. Even at low concentrations (0.5%) it can provide inhibition. It is, of course, toxic.
• Polyphosphates is the name of a family of polymers containing phosphate (—PO₄) groups. They are found in living organisms where they perform biological functions (DNA includes polyphosphates). They inhibit corrosion of most metals in water. Polyphosphates are not toxic, allowing their use in drinking water.
• Chromates are used mostly to inhibit corrosion of metals in nonacidic environments (a few, the alkali chromates, can be used to inhibit corrosion in acids; they are used to protect aluminum). They work by forming an adsorbed layer of chromate ions, passivating the surface. Chromates, like so many of these scary compounds, can be highly toxic, limiting their use.

Table 17.6 Corrosion inhibitors and the materials for which they work

Monitoring, maintenance and replacement

Regular inspection allows early indications of corrosion to be detected. Maintenance—painting, recoating or repair—can then be carried out, minimising down-time and risk of failure. There is a design issue here: the system must allow inspection of all potentially vulnerable surfaces, and permit access for restorative work. For safety-critical components a different strategy is adopted: that of replacement at prescribed, regular intervals, too short for corrosion to have caused any serious damage.