Chapter 3. History

Concrete is a construction material that is relatively easy to work with. However, concrete is very weak in tension in comparison to its compressive strength. Because of the low tensile strength of concrete, reinforcing steel bars are placed in regions of tension in a concrete member. This combination of concrete and steel provides a relatively inexpensive and durable material that has become widely used in construction of roadways and bridges. Reinforced concrete bridges have functioned reasonably well until the late 1960s, when premature concrete delamination and spalling, which used to be encountered only in coastal areas, became common in many of the reinforced concrete decks in the "snow belt" and concrete bridge decks were beginning to require maintenance after being in service for as little as 5 years. The emergence of this type of concrete deterioration, which was first observed in marine structures and chemical manufacturing plants, coincided with the increased application of deicing salts (sodium and calcium chlorides) to roads and bridges during winter months in those states where ice and snow are a problem to implement a "bare pavement policy".

It was recognized by the mid 1970s that this problem is caused by the corrosion of the reinforcing steel in the concrete which, in turns, is induced by the intrusion of even a small amount of chloride from the deicing salts into the concrete. It is difficult to estimate the cost of these corrosion-related damages to conventionally reinforced and prestressed concrete bridge components in the nation. According to a 1997 report, of the 581,862 bridges in and off the federal-aid system, about 101,518 bridges were rated as structurally deficient. Most of these bridges are not in danger of collapse, but they are likely to be load posted so that overweight trucks will be required to take a longer alternative route. The estimated cost to eliminate all backlog bridge deficiencies (including structurally and functionally) is approximately $78 billions (1), and it could increase to as much as $112 billions, depending on the number of years it takes to meet the objective. The average annual cost, through year 2011, for just maintaining the overall bridge conditions, i.e., the total number and the distribution of structurally and functionally deficient bridges, is estimated to be $5.2 billions. While corrosion of the reinforcing steel is not the sole cause of all structural deficiencies, it is a significant contributor and has therefore becomes a matter of major concern.

The magnitude of this corrosion problem in the transportation infrastructure has increased significantly in the last three decades and is likely to keep increasing. Even though the cost of maintaining bridge decks is becoming prohibitively expensive, the benefits provided by deicing salts are too great, however, that it's use is not likely to decrease in the future. In fact, the use of road deicing salts, which are extremely corrosive due to the disruptive effects of its chloride ions on protective films on metals, has actually increased in the first half of the 1990s—after a leveling off during the 1980s. Although an alternative effective and less corrosive deicing agent, calcium magnesium acetate (CMA), is available, its price is apparently not yet reasonable enough for winter maintenance engineers to use widely. Therefore, it can be expected that the road environment would likely remain corrosive, if not more, well into the future. In response to the tremendous economic burden that corrosion of reinforcing steel on concrete bridges placed on the national economy, the Structure Division of FHWA has placed emphasis on finding effective and economical solutions that can be easily implemented by various state and local transportation agencies.

In order to understand the various approaches by which this type of corrosion can be controlled, it is necessary to understand its nature. A few metals, notably gold, silver, and platinum, occur naturally. Engineering metals, including steel, must be derived from their ores by smelting. Ores are natural oxides, sulfides, and other reaction products of metals with the environment. During smelting, a metal absorbs the energy required to free it from the ore; and, this energy is retained within the metal after it is recovered. However, this metallic state is unstable, because the metal tends to rid itself of this extra energy by recombining with the environment to revert to its more stable and natural state as an ore. This reversion process is known as oxidation or, more specifically, corrosion.

A refined metal such as iron or steel has a natural tendency to corrode and thereby return to the stable state that it exists in nature, as iron ore (typically iron oxide, Fe₂O₃). The rate of steel corrosion depends on its composition, grain structure, and the presence of entrained stress from fabrication. It also depends on the nature of the surrounding environment, such as the availability of water, oxygen, and ionic species, pH and temperature.

In concrete, the presence of abundant amount of calcium hydroxide and relatively small amounts of alkali elements, such as sodium and potassium, gives concrete a very high alkalinity—with pH of 12 to 13. It is widely accepted that, at the early age of the concrete, this high alkalinity results in the transformation of a surface layer of the embedded steel to a tightly adhering film, that is comprised of an inner dense spinel phase (Fe₃O₄ / Fe₂O₃ ) in epitaxial orientation to the steel substrate and an outer layer of -FeOOH (2). As long as this film is not disturbed, it will keep the steel passive and protected from corrosion. When a concrete structure is often exposed to deicing salts, salt splashes, salt spray, or seawater, chloride ions from these will slowly penetrate into the concrete, mostly through the pores in the hydrated cement paste. The chloride ions will eventually reach the steel and then accumulate to beyond a certain concentration level, at which the protective film is destroyed and the steel begins to corrode, when oxygen and moisture are present in the steel-concrete interface.

In 1962, it was reported that the required minimum concentration of chloride in the concrete immediately surrounding the steel to initiate corrosion, the chloride corrosion threshold, is 0.15 percent soluble chloride, by weight of cement (3). In typical bridge deck concrete with a cement factor of 7, this is equivalent to 0.025-percent soluble chloride, by weight of concrete, or 0.59 kg soluble chloride per cubic meter of concrete. Subsequent research at FHWA laboratories estimated the corrosion threshold to be 0.033-percent total chloride, by weight of concrete (4,5). (Although it is widely accepted that only water-soluble (ionized) chloride contributes to corrosion, it is more common in practice to determined the total (inorganic) chloride contents of concrete samples from bridges, because analytical methods available for soluble chloride (AASHTO T-260 and ASTM C-1218) are more laborious and not as precise as that for total acid-soluble inorganic chloride (AASHTO T-260 and ASTM C-1152). In addition, it is common practice to express chloride contents in terms of weight percentage of concrete, which would not required an additional elaborate analysis to determine the cement content in the hardened concrete sample.) There are indications that the chloride corrosion threshold can vary between concrete in different bridges, depending on the type of cement and mix design used, which can vary the concentrations of tricalcium aluminate (C₃A) and hydroxide ion (OH^-) in the concrete. In fact, it has been suggested that because of the role that hydroxide ions play in protecting steel from corrosion, it is more appropriate to express corrosion threshold in terms of the ratio of chloride content to hydroxide content, [Cl^-] / [OH^-], which was recently established tote between 2.5 to 6 (6,7).

Once corrosion sets in on the reinforcing steel bars, it proceeds in electrochemical cells formed on the surface of the metal and the electrolyte or solution surrounding the metal. Each cell is consists of a pair of electrodes (the anode and its counterpoint, the cathode) on the surface of the metal, a return circuit, and an electrolyte. Basically, on a relatively anodic spot on the metal, the metal undergoes oxidation (ionization), which is accompanied by production of electrons, and subsequent dissolution. These electrons move through a return circuit, which is a path in the metal itself to reach a relatively cathodic spot on the metal, where these electrons are consumed through reactions involving substances found in the electrolyte. In a reinforced concrete, the anode and the cathode are located on the steel bars, which also serve as the return circuits, with the surrounding concrete acting as the electrolyte.

When corrosion occurs on the reinforcing steel in concrete, the electrochemical reactions involved are dependent on the environments at the steel-concrete interface:

• When Oxygen Is Present:
• 1. 
     
     

     Fe Fe+2 + 2 e-

     (1)

     
     
  2. At the anode, iron is oxidized to the ferrous state, releasing electrons.
     
  3. At the cathode, these electrons combine with oxygen and moisture to form hydroxide ions.
     
     

     ½ O2 + H2O + 2 e- 2 OH-

     (2)

     
     
  4. The ferrous ions combine with hydroxyl ions to produce ferrous hydroxide. Then the latter is further oxidized in the presence of moisture to form ferric oxide.
     
     

     Fe+2 + 2 OH- Fe(OH)2	(3)
     4 Fe(OH)2 + 2 H2O + O2 4 Fe(OH)3	(4)
     2 Fe(OH)3 Fe2O3 + 3 H2O	(5)

     
     
• When Oxygen is Absent:
  
• 1. 
     
  2. At the anode, in the upper layer of steel in a bridge deck, the oxidized iron reacts with chloride ions to form an intermediate iron complex.
     
     

     Fe+2 + 4 Cl- (FeCl4)-2 + 2 e-

     (6)

     
     
  3. This complex then reacts with moisture to form ferrous hydroxide.
     
     

     (FeCl4)-2 + 2 H2O Fe(OH)2 + 2 H+ + 4 Cl-

     (7)

     
     
  4. At the cathode, the hydrogen ions are reduced or combine with electrons to form hydrogen gas.
     
     

     2 H+ + 2 e- H2

     (8)

It has been suggested (8) that the chloride complex ions formed in reaction 6 may react with calcium hydroxide in the surrounding cement paste as follow:

(FeCl4)-2 + Ca(OH)2 Fe(OH)2 + CaCl2 + 2 CL-

(9)

It is apparent from the above reactions that, unfortunately, none of the chloride ions in the concrete are consumed and are, therefore, available again to contribute to corrosion. By forming hydrogen ions, or acid, in reaction 7, the pH at a local anodic site can reduce rapidly to values of 5 to 6. The resulting low-pH anode on a rebar is so different from that of other nearby rebars, which are surrounded by concrete of higher pH, that a powerful macro-cathode is created, which then feeds the original anodic spot. It is clearly obvious that moisture is required, not only to support the cathodic reactions but also to enhance the electrical conductivity of the concrete.

Corrosion can also occur even in the absence of chloride ions. For example, when the concrete comes into contact with carbonic acid resulting from carbon dioxide in the atmosphere, the ensuing carbonation of the calcium hydroxide in the hydrated cement paste leads to reduction of the alkalinity, to pH as low as 8.5, thereby permitting corrosion of the embedded steel:

CO2 + H2O H2CO3	(10)
H2CO3 + Ca(OH)2 CaCO3 +2 H2O	(11)

The rate of carbonation in concrete is directly dependent on the water/cement ratio (w/c) of the concrete, i.e., the higher the ratio the greater is the depth of carbonation in the concrete. In concrete of reasonable quality, that is properly consolidated and has no cracking, the expected rate of carbonation is very low. For example, in concrete with w/c of 0.45 and concrete cover 25 mm (1 in.), it will require more than 100 years for carbonation to reach the concrete immediately surrounding the steel (9). Carbonation of concrete or mortar is more of an issue Europe—thereby prompting the application of electrochemical realkalinization of concrete there—than in United States.

However, another possible source of problem to the durability of concrete that has not been widely recognized yet is the emissions of air pollutants such as sulfur dioxide (SO₂) and nitrous oxides (NO_x), resulting from burning of fossil fuels in the United States. When these gases come into contact with moisture in the atmosphere, they are converted to acids—in the form of acid rain and snow—that are considerably more corrosive than the carbonic acid resulting from carbon dioxide. These acids can cause acidification of cement paste that is more severe than the carbonization shown in reaction 11. Furthermore, there are indications that a combination of deicing salts and acid precipitation creates an even more corrosive environment than either of these substances alone does. In fact, there are data showing that corrosive areas in North America and Europe are, in general, areas where marine salts or deicing salts are present, while the most corrosive areas are those having a combination of these salts and acid precipitation (1). This synergistic effect of deicing salts and acids on metallic corrosion, which has been documented in laboratory studies, is attributed to the supply of cathodically reducible hydrogen ions for the corrosion reaction (10,11). It is conceivable that acid deposition from these gases can eventually have adverse effects on the concrete—increasing its permeability to intrusion by chloride ions and at the same time reducing the beneficially high alkalinity of the concrete around the reinforcing steel. The extent to which acid precipitation can become a threat to the durability of concrete deserves some investigations.

Each electrochemical corrosion cell exhibits a potential, or voltage, difference between the anode and the cathode, which drives the corrosion. In a reinforced concrete, the voltage difference may be created by a combination of the following:

• Differences in the surface of the steel bars. Steels are heterogeneous materials as they are alloyed with carbon and other elements. Their surfaces may be considered a patchwork of metal sites of slightly different electrochemical potentials. The presence of residual stress in the steel bars and even the presence of scratches on a portion of a bar and not on the other can create enough potential differences to drive corrosion.
• 
  
  
• Differences in the electrolytes. These can be differences in the concentration of chloride, oxygen, moisture, hydroxide, etc., in the concrete surrounding the steel bars. Such differences can readily exist in concrete, since it is a very heterogeneous material, both chemically and physically and on the microscopic and the macroscopic scales. Microscopic differences give rise to microscopic electrochemical cells (micro-cells), wherein each pair of electrodes (the anode and the cathode) exist in the same steel bar; in a concrete bridge deck, these paired electrodes would be found mostly in the upper reinforcing mat. Macroscopic differences create macroscopic electrochemical cells (macro-cells), wherein an anode is located on a steel bar in the upper mat and its cathode is located in another bars in the same mat or in a lower mat. The latter macro-cell is also common in bridge decks, where steel bars in the upper mat are anodic because of being exposed, in general, to higher amount of chloride ions and moisture than steel bars in the lower mat are.

The presence of cracks in the concrete can quickly give rise to these differences in different portions of the concrete surrounding the steel bars.

Lately it has been suggested that even heterogeneity in the distribution of porosity in the cement paste surrounding the reinforcing steel can affect surface corrosion of reinforcing steel—with corrosion reactions occurring preferentially on the surface of steel bars embedded in relatively denser hydrated cement paste (12). Measurements of surface impedance and phase angle, using a low-frequency impedance technique, revealed that depletion of oxygen at the steel/cement interface to form anodic areas may be the cause. It is no doubt that, in general, the greater these differences exist along the same steel bars or between neighboring steel bars, the faster is the resulting corrosion on the steel bars.

When corrosion on a steel bar progressed to the extent that the corrosion products (rust) occupy a greater volume than the steel and exert substantial stress on the surrounding concrete, the concrete begins to delaminate and then eventually spall. Estimates of the expansive force exerted on the concrete when steel converts to rust varied from 32 to 500 MPa (13,14). Similarly, different estimates have been made on the amount of corrosion product necessary to crack the concrete or mortar, and these have ranged from 0.1 to 20 mils (15,16). As corrosion takes place, the cross section of the steel is reduced, which leads to loss of bond between the steel and the concrete. This effect can be of serious concern in prestressed concrete bridge members, where the bonding between the high-strength tendons and the concrete is critical.

The rapid premature deterioration of many concrete bridge decks in the late 1960s had raised concern among the state highway agencies. As a result, the use of non-corroding fusion bonded epoxy-coated reinforcement as a corrosion protection system has become a standard practice since the late 1970s. For additional protection, low-permeability concrete (low water-cement ratio Portland cement concrete, latex- modified concrete, other specialty concrete, and improved mix design) and increased concrete cover over coated reinforcing steel began to be utilized. Use of waterproof membranes over concrete decks, in conjunction with an asphalt overlay, as a protection system has produced mixed results and is also used, although sparingly. Few states have used epoxy-coated rebars in conjunction with waterproof membrane and asphalt overlay as their preferred multiple corrosion protection system.

FHWA research and field data have indicated that the use of overlays, waterproof membranes, and sealers only serve to slow the corrosion rate but do not stop the ongoing corrosion process. These conventional rehabilitation methods marginally extend the life of a structure to a variable extent, depending upon the quality and the type of treatment employed. On the other hand, cathodic protection (CP) has proven to be successful in retarding and controlling chloride-induced corrosion in reinforced concrete (RIC) bridge components. In addition, the corrosion process can be stopped or slowed by the elimination of either chloride ions, moisture, or oxygen in the concrete. Electrochemical chloride extraction, which removes chloride ions from the contaminated concrete, is therefore another alternative rehabilitation technique that is being explored by a number of state highway agencies. More research is under way to bring this technology at par with cathodic protection.

Investigations are under way for development of protection systems that can be adopted in the construction of new reinforced concrete bridge substructures and prestressed concrete members (PS/C), such as piles, girders, box beams, etc., which are generally still being built with uncoated high-strength steel. Research in progress has shown promise for the use of epoxy-coated strands for PS/C structural concrete members. Another promising protective system consists of the use of a corrosion inhibitor as a concrete admixture during the mixing of concrete. Such admixture inhibits the oxidation of the reinforcing steel to the ferrous state and the subsequent formation of corrosion products. Corrosion inhibitors and low-permeability concrete have found extensive use in PS/C members in which black steel/strands are still generally used. However, as yet there is no generally accepted protection system for reinforced concrete substructure or prestressed concrete bridge members comparable to the acceptance of epoxy-coated reinforcement in concrete bridge decks.

The application of prestressed concrete technology in building bridges is relatively recent. Therefore, the existing prestressed concrete members in bridges are still relatively young, and the corrosion and the concrete deterioration problems associated with this type of concrete members have only became evident in the early 1980s. Although prestressed concrete members were generally manufactured with concrete of relatively higher strength, time has shown that they are subject to the same adverse effects of reinforcement corrosion as reinforced concrete members are. Documented cases of prestressed strands breaking as a result of corrosion make this a most pressing problem. Since PS/C members rely on the tensile strength of the strands to resist loads, loss of even a few strands per member could prove catastrophic. In addition, because of the high stress the strands are subjected to, corrosion effects are accelerated. Even small corrosion pits could cause fracture of a strand, as comp red to non-prestressed reinforcing steel that will literally rust away before breaking. Of the 581,862 bridges in the National Bridge Inventory, slightly more than 10 percent have prestressed concrete superstructures, all of which will eventually need some protective measures applied to them. In addition, many of the other structures not in the inventory have prestressed substructure members that will likewise need some degree of corrosion protection.

There is no doubt that corrosion of reinforcing steel in concrete bridges has become a costly problem for the nation's infrastructure. To address this, the Structures Division of the Federal Highway Administration has led the efforts to find solutions for preventing this problem in new constructions and for mitigating it in existing reinforced concrete bridges.