Groundwaters and soils containing sulfate salts may be liable to cause expansive sulfate attack. Sulfuric acid can also occur naturally in groundwaters and soils as a consequence of oxidative weathering of sulfide minerals such as pyrite and marcasite, both of which have the chemical formula FeS2. Oxidation to ferrous sulfate and sulfuric acid may arise in the presence of oxygen and moisture:
The reaction commonly stops here unless certain aerobic bacteria are present to produce further oxidation cycles (see Section 4.7). In undisturbed soils, the position of the water table is of great importance for determining access of air and therefore the extent to which oxidation can take place. If the pyrite is always below the water level, no oxidation occurs, whilst in areas of fluctuating water levels that permit replenishment of oxygen to the soil, the reaction proceeds readily. If there is sufficient calcium carbonate or other basic minerals in the soil, the sulfuric acid is neutralised as gypsum is formed. If not, as in some instances where sulfuric acid production is marked, the acid may persist in the groundwater (Eglinton, 1998).
Disturbance of natural ground conditions by excavation or drainage is a common cause of pyrite oxidation by providing access for air. Rapid oxidation has been reported when compressed air has been employed to keep back water during underground work, as in tunnel construction (Lea, 1970). This can result in large increases of acidity and in the sulfate content of groundwater. The oxidation of pyrite can be followed by other reactions that cause expansion, such as production of gypsum from any calcite present, which gives a solid volume increase of 103%, and the reaction of gypsum with the monosulfate-hydrocalumite solid solution in hardened concrete to re-form ettringite. The free sulfuric acid can also attack clay minerals like illite to form jarosite KFe3(SO4)2(OH)6, a hydrated potassium iron (III) sulfate that produces a volume 115% greater than that of pyrite. Ground heave can be caused adjacent to buried concrete structures by production of gypsum and other hydrated sulfates (Hawkins and Pinches, 1997; Cripps and Edwards, 1997). All these reactions can give rise to expansive forces within the concrete, which lower the long-term durability (Bensted, 1981a,b; Eglinton, 1998).
Ettringite is sometimes associated with thaumasite in aggressive ground when the temperatures are low (normally below ca. 15ëC) and either or both may be associated with expansion of the concrete if it arises (see Section 4.3). However, the mere presence of ettringite in a hardened concrete is not per se proof that it has produced expansion of the concrete. For instance, the ettringite which forms during the early stages of hydration of Portland cement does not cause expansion as it is not rigidly packed. Often, this ettringite is only partially converted to monosulfate-hydrocalumite solid solution (AFm) during the subsequent stages of cement hydration (Bensted, 2002c). Thus observation of ettringite, particularly by techniques such as SEM/EDXA spot analysis that sample only minute volumes within the specimens of concrete examined, clearly does not provide conclusive evidence of sulfate attack, as has been noted elsewhere (Bensted, 2002b; Neville, 2004).
Only a minority of the sulfate from the original gypsum ground in with the Portland cement clinker during manufacture ends up as ettringite, and most of its formation happens during the first day of hydration at normal ambient tempera- tures (Bensted, 1983). A substantial part of the sulfate derived from the gypsum actually ends up in the calcium silicate hydrate (C-S-H phase) either sorbed or in solid solution. Expansion of concrete takes place when the hardened mono- sulfate-hydrocalumite solid solution reacts with gypsum (either from external sources or still present in unhydrated parts of the cement along with unhydrated clinker minerals) to produce ettringite.