Mechanism of sulfate attack: a fresh look

Abstract The first paper in this two-part series [Cem. Concr. Res. 32 (2002) 915] summarized the experimental results from a comprehensive research study on sulfate attack. The current paper utilizes these results to develop models for the mechanism of attack resulting from sodium and magnesium sulfate solutions. Implications of changing the binder constituents or the experimental variables, such as concentration and temperature of the solution on the proposed mechanism, are also discussed. The potential of these mechanistic models for use in service life prediction models has also been identified. According to the proposed mechanism, the attack due to sodium sulfate solution progresses in stages. The expansion of an outer skin of the specimen leads to the formation of cracks in the interior region, which is chemically unaltered. With continued immersion, the surface skin disintegrates, and the sulfate solution is able to react with the hydration products in the cracked interior zone leading to the deposition of attack products in this zone. Now, this zone becomes the expanding zone, leading to further cracking of the interior of the mortar. In the case of magnesium sulfate solution, a layer of brucite (magnesium hydroxide) forms on the surface of the mortar specimen. The penetration of the sulfate solution then occurs by diffusion across this surface layer. As the attack progresses, the formation of attack products such as gypsum and ettringite in the paste under the surface leads to expansion and strength loss. The expansion also causes cracking in the surface brucite layer, and this leaves the mortar susceptible to direct attack by the magnesium sulfate solution. Conditions favorable for the decalcification of calcium silicate hydrate (C-S-H) are thus created, and the ultimate destruction of the mortar occurs as a result of the conversion of C-S-H to the noncementitious magnesium silicate hydrate (M-S-H).