Stages in Corrosion Induced Deterioration
Deterioration starts with the loss of protection provided by the concrete cover. This is followed by corrosion initiation and then propagation.The loss of protection provided by the concrete cover occurs as the result of the ingress of aggressive species such as chloride ions and carbon dioxide. These species are transported through the pore system in the cement paste. At low porosities, entrained and entrapped air voids are distinct cavities isolated from one another. This will restrict their influence on the transport of aggressive species. When they are not filled with solution they will act to block the solution transport of ions, while promoting transport of gaseous species. Cracks and other large voids may be oriented to produce a continuous network through the concrete thereby significantly reducing resistance to transport. If such large defects are absent, capillary pores will have the dominant effect on the transport of aggressive species into concrete.
The rate of any transport process will depend on the volume fraction, tortuosity and connectivity of the pores. This is determined by factors such as the water/cement ratio (w/c), cement content, cement fineness, cement type, use of cement replacement materials (for example ground granulated blastfurnace slag – GGBS, pulverised fuel ash - PFA and silica fume – SF), concrete compaction, and degree of hydration. Reducing w/c (which controls the original spacing of the cement grains) and prolonged hydration may, for example, result in the capillary pores becoming blocked by gel so that they are interconnected solely by gel pores. Transport is very slow through the gel pores.
It may be noted that no physical damage occurs while aggressive species contaminate the concrete cover. Indeed a gain in concrete strength may occur. Nevertheless the loss of protection provided by the concrete cover is a deterioration process that reduces the remaining maintenance-free service life.
The period of this first stage normally forms a substantial proportion of the service life before the first maintenance is necessary. Indeed it may account for more than 90% of the maintenance free service life of the concrete. This period is not only dependent on the properties of the cover concrete affecting the rate of transport of the aggressive species, but also on the cover depth. The thickness of the concrete cover may be as high as 60mm.
The second stage of the deterioration process involves corrosion propagation. The transition between the first and second stages is precipitated by the breakdown of the passive film resulting in corrosion initiation. During the second stage a loss of steel section occurs. Furthermore, because the high volume corrosion products exert tensile forces, spalling of the concrete cover may occur. This may affect the integrity of the structure by reducing the tensile and bond strength of the steel. In addition a safety hazard may result from the falling debris.
More detail on the stages in the deterioration process is given in the sections below.
Carbonation induced corrosion
When atmospheric carbon dioxide (CO2) dissolves in the cement pore solution, carbonic acid (H2CO3) is formed. A reduction in the pH of the concrete pore solution then occurs and some of the alkaline solid phases are neutralised. For example, calcium hydroxide (Ca(OH)2) will be converted into calcium carbonate (CaCO3). As the reserve levels of the alkaline solid phases are depleted, a zone of low pH (the carbonated zone) extends from the surface into concrete. This process is termed carbonation. Carbonation generally results in a small increase in the strength of the concrete. However a fall in pH to values below 10 at the steel may render the steel passive film thermodynamically unstable.
Factors associated with both the concrete and the external environment affect the rate of carbonation. The nature of the porosity and the alkaline reserves of the cement hydration products are the main factors associated with the concrete that affect carbonation. Thus, for example, a high water/cement ratio will increase the capillary porosity and the rate of carbonation. Carbonation induced corrosion at an early age in the life of a structure is often the result of low concrete cover.
Carbon dioxide reacts with all the major constituents of hydrated cement. This includes both calcium hydroxide and the calcium silicate hydrate gel. Many solid phases will release hydroxyl ions as the pH is reduced. These alkaline reserves in hydrated cement depend on the cement type. Table 2 shows the acid required to reduce the pH of aqueous suspensions of ground cement paste and concrete to a pH of approximately 10. It is apparent that blending ordinary Portland cement (OPC) with PFA or GGBS may substantially reduce the alkaline reserves in the hydrated paste and therefore reduce the resistance to carbonation.
An important environmental factor affecting the rate of carbonation is moisture. Water is necessary for the carbonation reaction, but if the pores are filled, the ingress of carbon dioxide is severely hindered. The moisture content of concrete is determined by the environment to which it is exposed. Thus concrete exposed in dryer environments carbonates more quickly.
Corrosion Initiation and Propagation
After the depth of carbonation has reached the reinforcing steel, the passive film is no longer stable. However significant rates of corrosion are not automatic. As indicated in Fig. 12, corrosion requires the presence of an electrolyte to conduct ions between the anodic and cathodic sites. This is normally provided by moisture in the concrete. In dry concrete the resistance of the environment to the flow of ionic current may be high. Increasing the moisture content increases the risk of significant corrosion occurring but it may reduce the carbonation depth.
The binding capacity for aggressive ions such as chloride is also affected by carbonation. Bound chloride ions are released as the pH of the concrete is reduced. This reduces the concrete resistivity and helps to retain moisture. Thus the presence of low levels of chlorides exacerbates the carbonation problem. Severe corrosion problems are encountered when high levels of chloride contamination accompany carbonation. However this is unusual as the dry conditions that promote high carbonation rates are not compatible with wet conditions that promote high rates of chloride contamination.
The depth of carbonation is shown by the indicator phenolphthalein. This changes colour when the pH decreases below 9.
The high resistivity of carbonated concrete coupled with the general breakdown of the passive film on steel in the carbonated zone results in the anodes and cathodes of the corrosion cell being closely spaced. Macro-cell corrosion is not important in carbonated concrete and a galvanic current is not required to maintain a change in the local in environment. Unlike chloride induced corrosion, carbonation induced corrosion is a problem that is relatively simple to solve.
Chloride Induced Corrosion
Chloride contamination of concrete may arise from both internal and external sources. Internal sources of chloride include contamination of the mix materials and the use of calcium chloride as a set accelerator in construction. Codes of practice now limit the acceptable levels of chloride contamination resulting from the use of contaminated mix materials, while the use of chloride containing admixtures for reinforced concrete is generally not permitted. Thus internal chlorides tend to affect older structures although accidental contamination of new structures may occur. For example the steel reinforcement in construction in a marine environment may be subjected to salt spray prior to casting the concrete and this may lead to corrosion induced cracking of the cover very early on in the life of the structure.
External sources of chlorides include de-icing salts and sea salt in marine environments. The ingress of such chloride into concrete occurs in the liquid phase. Chloride ions are transported through the pore solution in concrete by a number of mechanisms. These include diffusion due to a concentration gradient, migration in an electric field, and water flow. The rate of diffusion is described by the diffusion coefficient. This parameter gives the flux of a species (quantity passing through a unit area per unit of time) per unit of concentration gradient. Migration is determined by the ionic mobility. This gives the average velocity per unit of electric field. Water flow may result from a pressure gradient, absorption into partially dry concrete, wick action (absorption at one location with drying at another location) or electro-osmosis (the movement of water under the influence of an electric field). Each of these processes will have its own transport coefficient e.g. water permeability or water sorptivity.
Chloride transport is affected by the pore structure. The physical effect of the pore structure is generally incorporated into the measured values of the transport coefficients (e.g. the diffusion coefficient or ionic mobility) defining the various transport processes. In addition, chemical interactions between ions in the pore solution and with the pore walls affect the movement of chloride ions through concrete.
Two effects on chloride transport arising from interactions between ions in the pore solution are the effective concentration may differ from the actual concentration of a species and the separation of charge is constrained. The effective concentration of an ion resulting from its non-ideal behaviour is termed its activity. The theoretical description of the transport of a species may, to a good approximation, be expressed in terms of its concentration only when the concentration is low.
The constraint on charge separation implies that, when no ion source or sink is present, the net quantity of positive charge transported by cations must be equal to the negative charge transported by anions, after any initial transient effects have decayed. The transport of an excess of ions of one particular type (e.g. Cl-) in a given direction may only occur if other ions of the same sign (e.g. OH-) are transported in the opposite direction. Thus, when sodium chloride diffuses into concrete, chloride diffusion into the concrete might be retarded by its association with a slower charge balancing sodium ion if no significant counter diffusion of hydroxyl or sulphate ions out of the concrete occurs.
The main interactions between ions and the pore walls are chloride binding and membrane effects. Chloride binding in concrete may be defined as the interaction between the porous matrix and chloride ions which results in their effective removal from the pore solution. All cements bind a proportion of the chloride present. This will remove chloride from the transport process as well as alter the pore solution concentration and therefore the concentration gradient driving diffusion. While there are many factors associated with the constituents of the concrete, the composition of the pore solution and the external environment that affect chloride binding, the most important factor is the cement type.
A typical chloride binding isotherm for an ordinary Portland cement concrete is given in Fig. 19. This was derived from the analysis of empirical data extracted from a wide range of sources. It approximates a Langmuir binding isotherm. It may be noted that the chloride content of concrete in equilibrium with that in seawater equates to approximately 1.7% by weight of cement.
Ion exchange membranes are produced by a surface charge on the pore walls. This charge results from the incongruent dissolution of a net quantity of ions with the opposite charge. The charge on the pore walls is balanced by an equal and opposite charge in the pore solution. An ion exchange membrane therefore favours the transport of the charge balancing species through its pore system. For narrow pores or low concentrations, only charge balancing ions may be present in the pore solution. The sign and value of the surface charge will depend on the electrolyte composition and pH as well as on the binder type and composition of the hydration products. In the absence of membrane effects, the difference in the mobilities of the anions and cations will produce a potential difference termed a junction potential if diffusion occurs. When this is enhanced by membrane effects, it is termed a membrane potential.
The external environment also affects chloride ingress. High rates of contamination can occur in warm water saturated conditions. Severe corrosion induced deterioration has occuring just above the water line has been observed in a marine environment in the Middle East. At this location there will be some concentration of chloride ions as the result of evaporation from the concrete surface.
Chloride Induced Corrosion Initiation
The oxides that make up the passive film on iron are thermodynamically stable in the alkaline environment in concrete even when chloride ions are present . In this situation corrosion tends to be localised and chloride induced corrosion initiation in concrete follows the model of pitting corrosion. It is a two stage process involving pit nucleation and pit growth. The causes of pit nucleation are still subject to much debate. However pit nucleation is usually followed by repassivation. If the pits are to grow, pit nucleation must be accompanied by a local fall in pH and increase in chloride content at the pit nucleation site.
The local fall in pH occurs as the result of the hydrolysis of dissolved iron ions. An example of such a reaction is given by:
The presence of an excess of chloride ions provides the charge balancing anion to stabilise the local reduction in pH. Hydrochloric acid (HCl) is effectively formed (Fig. 21). Thus the continued dissolution of iron is sustained.
The effect of this mechanism is evident at voids at the steel-concrete interface. It was noted that corrosion tends to initiate at the location of such defects (Fig. 22). This has been attributed to the absence of solid phases on the steel at these locations that would otherwise release hydroxyl ions to prevent a local reduction in pH thereby inhibiting corrosion initiation. The first observations of this phenomenon were made on prestressing steel that was surrounded by a cement mortar. Difficulties in injecting the mortar into ducts containing the steel tendons resulted in defects that were the source of some catastrophic corrosion induced failures.
Titrametric methods have been developed to determine the resistance to pH reduction that gives rise to this inhibitive property of the solid phases. In one method, termed differential acid neutralisation analysis, the resistance to pH reduction is given by the acid added per unit of pH reduction of an aqueous suspension of ground solid. The dissolution of the solid phases gives rise to peaks in the data at various pH values. On the basis of such data, inhibitive effects have been attributed to many hydration products of cement that release hydroxyl ions into the pore solution at high pH values.
Chloride ions bound in the solid phase may act to promote corrosion via the same mechanism that operates when hydroxyl ions bound in the solid phase act to inhibit corrosion. The pH dependent dissolution characteristics of chloride are included in Fig. 23. It was noted that, for this sample, virtually all the chloride that was bound by the solid phases was released before the pH reduced to 11. As the passive film is still thermodynamically stable at this pH, it might be assumed that all the chloride released will be available to sustain pit growth. Thus solid phases may act as both inhibitors and aggressive agents in the corrosion of steel in concrete .
The chloride threshold level may be defined as the quantity of chloride at the steel that is necessary to sustain local passive film breakdown, and hence initiate the corrosion process. It is usually presented as a ratio of the total chloride to cement content of concrete (expressed as a weight percentage). Typical values range between 0.2 and 2.5% by weight of cement. While the chloride content is relatively easy to determine, it should be noted that the cement content is often only estimated as laboratory verification of this is more difficult.
Evidence of the participation of bound chloride in corrosion initiation comes from the absence of any significant correlation between chloride binding and chloride threshold level. Indeed, Fig. 24 indicates that the effect of defects at the steel-concrete interface has a much greater effect on the chloride threshold level than the cement type which controls the chloride binding capacity . However there is substantially more bound hydroxide than bound chloride which gives the solid phases a net inhibitive effect.
Low chloride threshold levels (0.2% Cl-) are often observed under site conditions. It is believed that these are the result of defects such as voids, millscale and rust on the steel. The presence of such defects combined with moisture (the electrolyte) and oxygen, which raises the steel potential to positive values, will reduce the amount of chloride that is necessary to cause passive film breakdown. Such defects are said to lower the pitting potential of the steel. This is the potential at which local breakdown of the passive film occurs. It generally shifts to more negative values as the chloride content increases. Because the condition of the steel-concrete interface is seldom characterised, the associated variations in the chloride threshold level appear to be random. Thus the chloride threshold level may be regarded as representing a risk of corrosion initiation. The risk of corrosion initiation occurring on UK bridges as a function of the chloride content may be estimated.
In addition to the condition of the steel-concrete interface, other factors may affect the chloride threshold levels to a smaller extent. These include factors associated with the external environment (moisture, temperature and chloride source), the barrier properties of the concrete cover (w/c and cover depth) and the chemical properties of the cementitious binder (pH buffering). Corrosion initiation may be prevented if the concrete is too dry, while in very wet environments the corrosion rate following initiation may be limited by the restricted access of oxygen. One of the effects of a dry environment may be to prevent moisture entering the entrapped air voids at the steel.