Concrete in Seawater



Concrete in Seawater

 

For several reasons, effect of seawater on concrete deserves special attention. First, coastal and offshore sea structures are exposed to the simultaneous action of a number of physical and chemical deterioration processes, which provide an excellent opportunity to understand the complexity of concrete durability problems in practice.

Second, oceans make up 80 percent of the surface of the earth; therefore, a large number of structures are exposed to seawater either directly or indirectly (e.g., winds can carry seawater spray up to a few miles inland from the coast). Concrete piers, decks, break-water, and retaining walls are widely used in the construction of harbors and docks. To relieve land from pressures of urban congestion and pollution, floating offshore platforms made of concrete are being considered for location of new airports, power plants, and waste disposal facilities. The use of concrete offshore drilling platforms and oil storage tanks is already on the increase.

 

Major
Ions
Concentration (mg/l)
Black
Sea
Marmara Sea
Mediterranean
Sea
North
Sea
Atlantic
Ocean
Baltic
Sea
Arabian
Gulf
BRE**
Exposure
Red
Sea
Sodium
4,900
8,100
12,400
12,200
11,100
2,190
20,700
9,740
11,350
Magnesium
640
1,035
1,500
1,110
1,210
260
2,300
1,200
1,867
Chloride
9,500
14,390
21,270
16,550
20,000
3,960
36,900
18,200
22,660
Sulfate
1,362
2,034
2,596
2,220
2,180
580
5,120
2,600
3,050
TDS
17,085
26,409
38,795
33,060
35,370
7,110
66,650
32,540
40,960
TDS Ratio*
3.90
2.52
1.72
2.02
1.88
9.37
1.00
2.05
1.63

Most seawater is fairly uniform in chemical composition, which is characterized by the presence of about 3.5 percent soluble salts by weight. The ionic concentrations of Na+ and Cl? are the highest, typically 11,000 and 20,000 mg/liter, respectively. However, from the standpoint of aggressive action to cement hydration products, sufficient amounts of Mg2+ and SO2? 4 are present, typically 1400 and 2700 mg/liter, respectively. The pH of seawater varies between 7.5 and 8.4, the average value in equilibrium with the atmospheric CO2 being 8.2. Under exceptional conditions (i.e., in sheltered bays and estuaries) pH values lower than 7.5 may be encountered; these are usually due to a higher concentration of dissolved CO2, which would make the seawater more aggressive to Portland cement concrete. The following Table shows the concentration of major ions in some of the world seas.

 

Concrete exposed to marine environment may deteriorate as a result of combined effects of chemical action of seawater constituents on cement hydration products, alkali-aggregate expansion (when reactive aggregates are present), crystallization pressure of salts within concrete if one face of the structure is subject to wetting and others to drying conditions, frost action in cold climates, corrosion of embedded steel in reinforced or prestressed members, and physical erosion due to wave action and floating objects.

Attack on concrete due to any one of these causes tends to increase the permeability; not only would this make the material progressively more susceptible to further action by the same destructive agent but also to other types of attack. Thus a maze of interwoven chemical as well as physical causes of deterioration are found at work when a concrete structure exposed to seawater is in an advanced stage of degradation. The theoretical aspects, selected case histories of concrete deteriorated by seawater, and recommendations for construction of concrete structures in marine environment are discussed next.

 

 

Theoretical Aspects

From the standpoint of chemical attack on hydrated Portland cement in unreinforced concrete, when alkali reactive aggregates are not present, one might anticipate that sulfate and magnesium are the harmful constituents in seawater. It may be recalled that with groundwater, sulfate attack is classified as severe when the sulfate ion concentration is higher than 1500 mg/liter; similarly, Portland cement paste can deteriorate by cat-ion-exchange reactions when magnesium ion concentration exceeds, for instance, 500 mg/liter.

Interestingly, in spite of the undesirably high sulfate content of seawater, it is a common observation that even when a high C3A Portland cement has been used and large amounts of ettringite are present as a result of sulfate attack on the cement paste, the deterioration of concrete is not characterized by expansion; instead, it mostly takes the form of erosion or loss of the solid constituents from the mass. It is proposed that ettringite expansion is suppressed in environments where (OH)? ions have essentially been replaced by Cl? ions. Incidentally, this view is consistent with the hypothesis that alkaline environment is necessary for swelling of ettringite by water adsorption.

Irrespective of the mechanism by which the sulfate expansion associated with ettringite is suppressed in high C3A Portland cement concretes exposed to seawater, the influence of chloride on the system demonstrates the error too often made in modeling the behavior of materials when, for the sake of simplicity, the effect of an individual factor on a phenomenon is predicted without sufficient regard to the other factors present, which may modify the effect significantly.

It may be noted that according to ACI Building Code 318-83, sulfate exposure to seawater is classified as moderate, for which the use of ASTM Type II Portland cement (maximum 8 percent C3A) with a 0.50 maximum water/cement ratio in normal-weight concrete is permitted. In fact, it is stated in the ACI 318R-21 Building Code Commentary that cements with C3A up to 10 percent may be used if the maximum water/cement ratio is further reduced to 0.40.

 

The fact that the presence of uncombined calcium hydroxide in concrete can cause deterioration by an exchange reaction involving magnesium ions was known as early as 1818 from investigations on disintegration of lime-pozzolan concretes by Vicat, who undoubtedly is regarded as one of the founders of the technology of modern cement and concrete. Vicat made the profound observation:

 

On being submitted to examination, the deteriorated parts exhibit much less lime than the others; what is deficient then, has been dissolved and carried off; it was in excess in the compound. Nature, we see, labors to arrive at exact proportions, and to attain them, corrects the errors of the hand which has adjusted the doses. Thus the effects which we have just described, and in the case alluded to, become the more marked, the further we deviate from these exact proportions.

 

Several state-of-the-art reviews on the performance of structures in marine environments confirm that Vicat’s observation is equally valid for Portland cement concrete. From long-term studies of Portland cement mortars and concretes exposed to seawater, the evidence of magnesium ion attack is well established by the presence of white deposits of Mg(OH)2, also called brucite, and magnesium silicate hydrate. In seawater, well-cured concretes containing large amounts of slag or pozzolan in cement usually outperform reference concrete containing only Portland cement, partly because the former contain less uncombined calcium hydroxide after curing. The implication of loss of lime by cement paste, whether by magnesium ion attack or by CO2 attack, is obvious from Figure 1.

Since seawater analyses seldom include the dissolved CO2 content, the potential for loss of concrete mass by leaching away of calcium from hydrated cement paste due to carbonic acid attack is often overlooked. According to Feld, in 1955, after 21 years of use, the concrete piles and caps of the trestle bends of the James River Bridge at Newport News, Virginia, required a $ 1.4 million repair and replacement job which involved 70 percent of the 2500 piles. Similarly, 750 precast concrete piles driven in 1932 near Ocean City, New Jersey had to be repaired in 1957 after 25 years of service; some of the piles had been reduced from the original 550 mm diameter to 300 mm. In both cases, the loss of material was associated with higher than normal concentrations of dissolved CO2 present in the seawater.

It should be noted that in permeable concrete the normal amount of CO2 present in seawater is sufficient to decompose the cementitious products eventually. The presence of thaumasite (calcium silicocarbonate), hydrocalumite (calcium carboaluminate hydrate), and aragonite (calcium carbonate) have been reported in cement pastes derived from deteriorated concretes exposed to seawater for long periods.

 

Figure 1. Strength loss in permeable concrete due to lime leaching (adopted from I. Biczok, Concrete Corrosion and Concrete Protection, Chemical Publishing Company, Inc., New York, 1967, p.291.

 

Lessons from Case Histories

 

For the construction of concrete structures in marine environment, important lessons from case histories of concrete deteriorated by seawater can be summed up as follows:

 

1. Permeability is the key to durability. Deleterious interactions of serious consequence between constituents of hydrated Portland cement and seawater take place when seawater is not prevented from penetrating into the interior of a concrete. Typical causes of insufficient watertightness are poorly proportioned concrete mixtures, absence of properly entrained air if the structure is located in a cold climate, inadequate consolidation and curing, insufficient concrete cover on embedded steel, badly designed or constructed joints, and microcracking in hardened concrete attributable to lack of control of loading conditions and other factors, such as thermal shrinkage, drying shrinkage, and alkali aggregate expansion. It is interesting to point out that engineers on the forefront of concrete technology are becoming increasingly conscious of the significance of permeability to durability of concrete exposed to aggressive waters. For example, concrete specifications for offshore structures in Norway now specify the maximum permissible permeability directly (k ? 10?13 kg/Pa· m · sec).

 

2. Type and severity of deterioration may not be uniform throughout the structure (Figure 2). For example, with a concrete cylinder the section that always remains above the high-tide line will be more susceptible to frost action and corrosion of embedded steel. The section that is between high- and low-tide lines will be vulnerable to cracking and spalling, not only from frost action and steel corrosion but also from wet-dry cycles. Chemical attacks due to alkali-aggregate reaction and seawater cement paste interaction will also be at work here. Concrete weakened by microcracking and chemical attacks will eventually disintegrate by action and the impact of sand, gravel, and ice; thus maximum deterioration occurs in the tidal zone. On the other hand, the fully submerged part of the structure will only be subject to chemical attack by seawater; since it is not exposed to subfreezing temperatures there will be no risk of frost damage, and due to lack of oxygen there will be little corrosion. It appears that progressive chemical deterioration of cement paste by seawater from the surface to the interior of the concrete follows a general pattern.

The formation of aragonite and bicarbonate by CO2 attack is usually confined to the surface of concrete, the formation of brucite by magnesium ion attack is found below the surface of concrete, and evidence of ettringite formation in the interior shows that sulfate ions are able to penetrate even deeper. Unless concrete is very permeable, no damage results from chemical action of seawater on cement paste because the reaction products (aragonite, brucite, and ettringite), being insoluble, tend to reduce the permeability and stop further ingress of seawater into the interior of the concrete. This kind of protective action would not be available under conditions of dynamic loading and in the tidal zone, where the reaction products are washed away by wave action as soon as they are formed.

 

3. Corrosion of embedded steel is, generally, the major cause of concrete deterioration in reinforced and prestressed concrete structures exposed to seawater, but in low-permeability concrete this does not appear to be the first cause of cracking. Based on numerous case histories, it appears that cracking corrosion interactions probably follow the route diagrammatically illustrated in Figure 3. Since the corrosion rate depends on the cathode/anode area, significant corrosion and expansion accompanying the corrosion should not occur until there is sufficient supply of oxygen at the surface of the reinforcing steel (i.e., an increase in the cathode area).

This does not happen as long as the permeability of steel-cement paste interfacial zone remains low. Pores and micro cracks already exist in the interfacial zone, but their enlargement through a variety of phenomena other than corrosion seems to be necessary before conditions exist for significant corrosion of the embedded steel in concrete. Once the conditions for significant corrosions are established, a progressively escalating cycle of cracking corrosion- more cracking begins, eventually leading to complete deterioration of concrete.

 

 

 

 

Figure 2. Diagrammatic representation of deterioration of concrete cylinder exposed to seawater.

 

 

 

 

Figure 3. Diagrammatic representation of the cracking-corrosion-cracking cycles in concrete.