Sulfate Attack and Sulfate Content in Concrete: A Literature Review

Ingress and transport of sulfate through the pore system of concrete results in the chemical reaction between sulfate and cement hydration product phases. Creation of expansive reaction products generates high stress within the pore system, which ends up in degradation of cementitious systems with a mechanical response as spalling and cracking. Some products are not expansive but are soft non-binders. The sulfate compounds responsible for sulfate attack are water soluble salts of alkaline- earth metals (calcium and magnesium) and alkali metals (sodium and potassium). Apart from chemical attack, sulfate salts can attack concrete by a physical mechanism like cyclic crystallization and dissolution or hydration and dehydration with mineral phase change.

ACI's Guide to Durable Concrete (201.2R) identifies two primary mechanisms of sulfate attack- the formation of ettringite and the formation of gypsum. But in detail, five different types of chemical reactions of sulfate with cement hydration products have been identified. Acid attack by sulphuric acid is not included in this literature research work.

The most important one is the reaction with hydrated aluminate phase (C3A) to form the expansive compound ettringite, the primary destructive compound in sulfate attack.

Secondly, sulfate ion reacts with calcium hydroxide (Ca(OH)2) in the paste to form gypsum (CaSO4. 2H2O).

Another type of reaction is by magnesium sulfate (MgSO4) with calcium silicate hydrate (C-S-H) phase. 

In the fourth type of reaction, carbonate also involved along with sulfate. The calcium silicate hydrate (C-S-H) and calcium hydroxide (Ca(OH)2) are decomposed as a result of reaction with sulfate and carbonate to form thaumasite (CaSiO3.CaCO3.CaSO4.15H2O).

The last type of reaction is the crystallization of sodium sulfate (Na2SO4) salt due to wetting and drying, which attacks concrete and appears as surface scaling. Currently, this type of sulfate attack has been classified as Physical Sulfate Attack (PSA).

Sulfate attacks are categorized into two broad classes based on the origin of source of sulfate, as either external sulfate attack (ESA) or internal sulfate attack (ISA). In external sulfate attack, the source of sulfate is external to concrete and often originate from its surrounding like soil, sea water, sewerage water, groundwater and the like. Sulfate ions contained in the concrete ingredients like cement, water, aggregate and admixture are the reason for internal sulfate attack.

External sulfate attack is the most common type and typically occurs where water containing dissolved sulfate penetrates the concrete. The composition and microstructure of concrete will change due to this sulfate reaction. The net effects are extensive cracking, expansion and loss of bond between the cement paste and aggregate. Concrete should be designed to have low permeability and porosity to resist such external sulfate attack.

 Fig. 1- Types of sulfate attack

Internal sulfate attack occurs when a source of sulfate is incorporated in the concrete when mixing like excess gypsum in cement, sulfate in the aggregate, water or admixture, and contamination. Proper screening and testing procedures should be used to avoid internal sulfate attack.

Ettringite Formation

This is the ‘classic and conventional’ kind of sulfate attack. Ettringite is one of the first group of mineral phases formed in hydrated Portland cement paste system as a result of the reaction of calcium aluminate and calcium sulfate. Naturally occurring ettringite is a hydrous calcium aluminium sulfate mineral with formula Ca6Al2(SO4)3(OH)12. 26H2O. The name was after the type locality at Ettringen, Rhineland, Germany. In concrete chemistry, ettringite is a hexacalcium aluminate trisulfate hydrate, of general formula (CaO)3(Al2O3)(CaSO4)3. 32H2O.

 

 

Fig. 2- Naturally occurring ettringite crystal

 

The hydration of calcium aluminate (C3A) and calcium alumino ferrite (C4AF) phases of Portland cement is somewhat more complex than that of the calcium silicate (C3S and C2S) phase. The reactions depend on whether sulfate ions are present in the pore solution. C3A is very soluble, and the reaction is very rapid that if it is allowed to occur it would release large amounts of heat and could cause the paste to set within a few minutes after mixing (flash set). The purpose of adding calcium sulfate (CaSO4) in Portland cement is to prevent this rapid reaction. Calcium sulfate can be gypsum (CaSO4. 2H2O), hemihydrate (CaSO4. ? H2O) or anhydrite (CaSO4) or any mixture of them (BS EN 197-1).The presence of sulfate ions from calcium sulfate (CaSO4) causes the C3A to undergo a different hydration reaction to form ettringite.

C3A + 3(CaSO4. 2H2O) + 26 H2O ---> C3A. 3CS. 32H 

 

 

Fig. 3- Ettringite (white, needle-like crystals)

Thus, ettringite is formed within the first few hours after mixing cement with water. All of the sulfur in the cement is generally consumed to form ettringite within twenty-four hours. This process is termed as Early Ettringite Formation (EEF) and the ettringite thus formed is called primary ettringite, which does not cause any significant localized disruptive action even though it has a higher volume than its reactants. As concrete is still in the plastic stage, this expansion is harmless and often un-noticed. At this stage, ettringite is uniformly and discretely dispersed throughout the cement paste at a sub-microscopic level.

Primary ettringite is responsible for the setting regulation in concrete. The retardation has been attributed to the early formation of an ettringite layer which acts as a coating over the surfaces of cement grains soon after mixing. The classical view that ettringite does not contribute to strength and durability cannot be true. Data taken in isolation are not necessarily relevant. Primary ettringite can improve strength, reduce permeability and porosity, and can give dimensional stability due to its pore filling effect. 

At high curing temperatures over 65- 70C, the primary ettringite formed will be decomposed, and the sulfate and alumina are held in the C-S-H gel of the cement paste or partially converted to monosulfate (AFm) and sulfate ions. As the temperature of concrete falls below 70C, the ettringite subsequently re-forms causing expansion in the concrete which has already been hardened. This process is called Delayed Ettringite Formation (DEF) and the ettringite thus formed is called secondary ettringite. The volume stress due to delayed ettringite formation leads to cracking. However, M. Collepardi’s hypothesis is against this school of thought and is based on the role played by the late sulfate release (from cement and aggregate) on delayed deposition of ettringite in the pre-existing micro cracks. In fact, secondary ettringite can form in both ways- high curing temperatures above 70C and late sulfate release from slow soluble sulfate minerals.

The hypothesis for a deterioration mechanism related slow soluble sulfates is that some portions of the sulfate in clinker are in a form that is very slowly soluble such as β-anhydrite. Moreover, if concrete is exposed to water for extended periods of time, primary ettringite can slowly dissolve and reform in any available voids or micro cracks.

Delayed ettringite formation has been reported in concrete which has been cured at elevated temperatures (steam curing) and in large pours where the heat of hydration has resulted in high temperatures (adiabatic temperature rise) within the core area. DEF causes the paste to expand while the aggregate does not, which results in cracks (or gaps) around the aggregate: the bigger the aggregate, the more significant the gap. Often these gaps become filled with ettringite. The strength is lost as the aggregate are effectively detached from the cement paste framework. DEF can increase the risk of the secondary form of deterioration such as reinforcement corrosion due to the ingress of hazardous chloride ions through the DEF generated cracks.

BS 8500 Part 1: 2006 warns about DEF and specifies (clause A.8.2) where the heat of hydration or accelerated curing is likely to take the concrete temperature above 70C, the potential for DEF should be considered. It would be worth to note in this context that BS 8110 Part 2: 1985 clause 3.8.4.1 limits temperature differential to 20C to avoid cracking due to a thermal gradient.

 

 

Fig. 4- Delayed ettringite formation: gap around limestone aggregate.

Limiting the internal concrete temperature below 70OC during its early life will minimize the risk of subsequent DEF. This can be achieved by limiting the cement content or by the use of low heat cement. M. Collepardi, S. Collepardi, J.J. Ogoumah Olagot and R. Troli have reported, from a practical point of view, with a curing at temperature lower than 80OC (preferably lower than 70OC) there is no thermal decomposition of ettringite and then no risk of DEF related damage in concrete structures independently of the SO3 content in the clinker phases (up to 2.7%) or Portland cement (up to 4.2%). However, if, for some reasons, a curing temperature higher than 80OC, preferably 70OC, cannot be excluded, then it would be better to limit the amount of SO3 in the cement at a percentage lower than 4.0 by using, for instance, blended cement. 

Gypsum Formation

Chemically known as ‘calcium sulfate dihydrate', Gypsum is a soft (2 in Mohs' scale of hardness) sulfate mineral with the chemical formula CaSO4. 2H2O. It is widely used as fertilizer and in cement industry to prevent the flash set of due to the rapid reaction of C3A with water. In the cement manufacturing process, upon the cooling of clinker, a small amount of gypsum (typically 3 to 10%) is added during the final grinding.

 

 

Fig. 5- Gypsum mineral

The hydration of silicate phases of Portland cement releases lime as Ca(OH)2. The sulfate ions react with calcium hydroxide to form gypsum. This reaction product also has a greater solid volume than the original constituents and in some cases can contribute to the degradation of the concrete.

The reactions of sodium sulfate and magnesium sulfate with calcium hydroxide are as follows:

Na2SO4+Ca(OH)2 +2H2O --- > CaSO4.2H2O +2NaOH

MgSO4 + Ca(OH)2 + 2H2O --- > CaSO4.2H2O + Mg(OH)2

Brucite Formation

Naturally occurring brucite was first described in 1824 and named after the discoverer, American Mineralogist Archibald Bruce (1777 – 1818).

 

 

Fig. 6- Naturally occurring brucite crystal

 

Similar to ettringite and gypsum, brucite also is an expansive mineral, the formation of which induces mechanical stress in the hardened cement paste. Magnesium sulfate has two-fold reaction with cement paste and hence is generally more aggressive, at the same concentration. The magnesium also takes part in the reaction replacing calcium in the solid phase apart from its reaction with Ca(OH)2 to form brucite (Mg(OH)2). The replaced calcium precipitates mainly as gypsum. Consequently, the calcium silicate hydrate (C-S-H) is converted to magnesium silicate hydrate (M-S-H).

The mechanism of magnesium sulfate attack with cement hydrates are two-fold as follows:

MgSO4 + Ca(OH)2 --- > CaSO4. 2H2O + Mg(OH)2

MgSO4 + C-S-H --- > CaSO4. 2H2O + M-S-H

The de-dolomitization of reactive dolomite aggregate ((CaMg)2 CO3) can fuel up the formation of brucite.

Thaumasite form of Sulfate Attack (TSA)

Thaumasite is a silicate mineral with chemical formula CaSiO3.CaCO3.CaSO4.15H2O. It occurs as hydrothermal alteration mineral in sulfide mines, carbonate rocks, more rarely basaltic rocks. It was first described in 1878 in Sweden. The name was derived from the Greek word ‘thaumazein', which means ‘to be surprised,' because of its remarkable composition, without parallel at the time of its discovery.

 

 

Fig. 7- Naturally occurring thaumasite crystals. 

 

Considerably distinct from conventional ettringite form of sulfate attack, where calcium aluminate and sulfate ions are required for the reaction, thaumasite form of sulfate attack requires carbonate as well as sulfate ions in solution. The calcium silicate hydrate (C-S-H) and calcium hydroxide (Ca(OH)2) are decomposed as result of reaction with sulfate and carbonate to form thaumasite.

 

Fig. 8- Thaumasite formation in concrete.

TSA tends to form at low temperatures about 4 to 10OC, and significant damage only occurs in a very wet ground. TSA is a relatively rare form of sulfate attack, and risk of its occurrence is thought to be negligible unless the following conditions are met (Thaumasite Expert Group (TEG) - 1999)

1. Low temperature (generally below 15OC)

2. Presence of sulfate ions

3. Presence of carbonates

4. Presence of mobile ground water

It has been shown that thaumasite may form if concrete containing commonly used limestone aggregates are exposed to sulfate solutions. As it forms, the concrete converts to a friable material. Concrete severely affected by TSA can easily be broken with fingers and the coarse aggregate is taken out. Since it is the calcium silicate hydrate in concrete that provides most of the strength, TSA results in severe weakening of concrete. TSA can completely destroy the binding property of cementitious materials of concrete by transforming it into a mush.

 Fig. 9- Concrete severely affected by TSA

Many types of research have been done on the use of limestone as a component of Portland cement, particularly in relation to the risk of sulfate attack due to thaumasite formation. Up to 5% limestone in Portland cement does not appear to be any significantly increased susceptibility to TSA. Research does exists concerning much higher levels of limestone (15 to 35%, or where the carbonate fines originate from the aggregate), where used in cold temperatures combined with wet and aggressive sulfate environments, that indicates more susceptibility to the thaumasite form of sulfate attack (R. Doug Hooton and Michael D.A. Thomas).

Physical Sulfate Attack (PSA)

Physical sulfate attack is often evidenced by bloom (presence of sodium sulfate salts) at exposed concrete surface. It is not only a cosmetic problem but the visible display of possible chemical and microstructural degradation within the concrete. 

Fig. 10- Physical sulfate attack.

Sulfate salts enter the pore spaces of concrete in solution by capillary action and diffusion. The wick's action brings the sulfate solution to the exposed surface, and by evaporation, the sulfate ion concentration increases gradually until it crystallizes. Changes in ambient temperature and relative humidity cause the sulfate salts to undergo cycles of crystallization and dissolution, or hydration and dehydration. When crystallization and hydration are accompanied by volumetric expansion, repeated cycles can cause deterioration of concrete similar to that caused by cycles of freezing-and-thawing.

The phase change of sodium sulfate salts like mirabilite (Na2SO4. 10H2O – decahydrate hydrous form) and thenardite (Na2SO4 – anhydrous form) causes a huge difference in crystallization pressure within the pore spaces. At a relatively high temperature, the anhydrous thenardite crystallizes from saturated pore solution and subsequent conversion to hydrated mirabilite requires substantially greater volume. The wet and humid growth of salts in pore spaces and the consequently induced stresses can overcome the tensile strength of concrete. Such stresses are directly related to the crystallization pressure. Crystallization pressure is the pressure necessary for the stability of a crystal in its supersaturated solution (Francesco Caruso and Robert J. Flatt). 

Protective Measures

High cement content, more than 385 kg/m3, with corresponding low water-cement ratios, are very beneficial to sulfate resistance in general, where the basic concept is to produce concrete with lower porosity and permeability as a primary defensive measure against ingress of sulfate ion. Sulfate resistant portland cement (SRC) provides some defense against ettringite type sulfate attack as it has low calcium aluminate, but it does not give any particular protection against thaumasite formation. Concrete made with good quality carbonate aggregate, and GGBS performed well in the long term than that made with SRC (M.A.Halliwell and N.J.Crammond). Lindon K.A. Sear had shown that increasing the fly ash content ≥36% seems to enhance the resistance to thaumasite considerably. He assumes this could be simply a dilution effect. Silica fume also can offer better resistance to sulfate attack. S.T.Lee, H.Y.Moon and R.N.Swamy had reported from their silica fume research studies that the best resistance to sodium sulfate attack was obtained with an SF replacement of 5 to 10% replacement, but even then, a strength loss of 15 to 20% could be expected. On the other hand, mortars with silica fume were severely damaged in magnesium sulfate environment. ). P.M. Carmona-Quiroga and M.T. Blanco-Varela had shown that addition of (5 to 20% by weight) BaCO3 to Portland cement raised its thaumasite resistance. The effect can be attributed to the formation of BaSO4, highly insoluble phase that immobilizes part of the external sulfates.

Sulfate attack by gypsum formation can be reduced by the use of blended cement containing blast furnace slag or pozzolan. These supplementary cementitious materials react with Ca(OH)2 to form additional C-S-H, so that Ca(OH)2 is no longer available for reaction with sulfates. Moreover, it results in the reduction in Portland cement per cubic meter of concrete and hence less Ca(OH)2. It is important to note that Class C fly ash decreases the sulfate resistance of concrete because of its high calcium oxide content along with alumina. BS 5328-1 specifies GGBS with alumina content greater than 14% should be used only with Portland cement having a C3A content not exceeding 10%. 

Unlike chemical sulfate attack, where standards have been developed to counter it, there are no official standards for mitigating physical sulfate salt attack. But there are ways of handling it, such as isolating the rising damp using slick bands that interfere with its climbing prowess and by keeping water away. Air- entrained mixes at the same w/cm performed better than non- air mixes as air- entrainment can provide space for salt as well as it creates capillary breaks (R. Doug Hooton).

The factors that mitigate or delay sulfate (chemical) attack in concrete can be (ACI 201.2R- Guide to Durable Concrete):

1- Dense concrete achieved by proper mix design, low w/c ratio, increased cement content, air entrainment, proper consolidation, and adequate curing.

2- Reduced tensile stress in concrete by using adequate tensile reinforcement, the inclusion of pozzolan, provision of appropriate contraction joints. 

3- Proper structural design to minimize the area of contact and turbulence and provision of membranes and protective barrier system to reduce penetration.

Specification Limits

Specifications for many constituent materials (cement, GGBS, fly ash, silica fume, water, and aggregates) place limits on the sulfate content. The majority of the standards specify sulfate as SO3, but some standards quote sulfate as SO4. However, the conversion factor is SO3 = SO4 x 0.833. 

Clause 6.2.5.3 of BS 8110 Part-1: 1995, Structural use of concrete had a limit of 4% by mass of cement based on the total acid soluble content of the concrete mix, expressed as SO3. This restriction was dropped in the 1997 edition. BS 5328-1 suggests, as no tests exist to determine mobile sulfate content, it is usual to measure the acid soluble sulfate contents of the constituents. The relationship between such measurements and the mobile sulfate content in the hardened concrete is variable, and therefore no universal sulfate limit can sensibly be applied to concrete.

The recommendations for concrete (dense, fully compacted made with 20mm nominal maximum size aggregate conforming to BS 882 or BS 1047) exposed to sulfate attack in accordance with BS 5328-1 can be summarized as follows.

The minimum cement content specified in the mix shall be increased by 40 kg/m3 for concrete made with 10mm nominal maximum aggregate size. The cement group mentioned in the above table can be summarized as given below.

The BS 5328 series had been withdrawn in 2003, and the English version of EN 206-1 supersedes them. In the UK, the BS EN 206-1 shall be used in conjunction with the complimentary standards BS 8500-1 and BS 8500-2.

In BS EN 206-1, the concrete exposed to chemical attack has been classified under XA Class, where the presence and concentration of SO4, CO2, NH4, and Mg are considered along with pH. The aggressive chemical environment classified in this standard are based on natural soil and ground water at water/soil temperatures between 5OC and 20OC and a water velocity sufficiently slow to approximate to static conditions.

 The BS 8500-1 refers to ACEC Classes (Aggressive Chemical Environment for Concrete) rather than the XA Classes used in BS EN 206-1. 

The sulfate classes stated in BS 5328-1 are similar to the sulfate classes given BS 8500-1 and BRE Special Digest 1, where they are identified as DS1, DS2, DS3, DS4, DS4m, DS5 and DS5m with total potential sulfate percentage (BS 8500-1: 2006, Table A.2 and BRE Special Digest-1, Table C1). This five level Design Sulfate Class (DS Class) is for sites, based on sulfate content of the ground and/or ground water.

The DS Class for the site shall be determined by carrying out the ground investigation which includes ground water mobility (static, mobile, and flowing), sulfates, sulfides, magnesium and acids (pH). Based on DS Class, type of site (natural or brownfield), water mobility and pH, the Aggressive Chemical Environment for Concrete (ACEC) Class shall be determined (BS 8500-1: 2006, Table A.2 and BRE Special Digest-1, Table C1). The Design Chemical Class (DC Class) is derived from the ACEC Class and takes into account a number of other factors including intended working life (BRE Special Digest-1, Table D1 and D2).

Many key changes have been made to the procedure for concrete specification in the third edition of BRE SD-1: 2005. The SO4 levels in the revised version of DS Class has considerably changed. Some soil that was previously classified as DS1 would now be considered as DS3. The new limits bring SO4 classification based on 2:1 water/ soil extract tests into parity with the ground water based tests. Other changes include maximum w/c ratio, minimum cement content and the number of APM (Additional Protective Measures) to be applied. The use of the concept ‘intended working life’ has replaced that of ‘structural performance level.' 

The American method is comparatively much simple. The Category S in ACI 318- Building Code Requirement for Structural Concrete applies to concrete in contact with soil or water containing deleterious amounts of water soluble sulfate ions. Representative samples of water and soil shall be tested to assess the severity of the potential exposure of concrete to detrimental amounts of sulfate. ACI 318 identifies four exposure classes for sulfate as S0, S1, S2 and S3 (Table 4.2.1.) and describes the concrete design requirements to protect against damage to concrete by sulfate attack (Table 4.3.1).

References

A.M. Neville- Properties of Concrete, Pearson Education (Singapore).

ACI 201.2R- Guide to Durable Concrete.

ACI 318- Building Code Requirements for Structural Concrete and Commentary.

BRE Special Digest-1: 2005- Concrete in Aggressive Ground.

BS 5328-1:1997- Guide to Specifying Concrete.

BS 8110-1:1997- Code of Practice for Design and Construction.

BS 8500-1:2015- Method of Specifying and Guidance for the Specifier.

BS EN 197-1:2011- Composition, Specifications and Conformity Criteria for Common Cements.

BS EN 206: 2013- Concrete: Specifications, Performance, Production, and Conformity.

Many other published research papers on sulfate attack in concrete.