User:Shinkolobwe/Cement swelling phases

Cement swelling phases are expansive minerals which can induce the formations of cracks in hardened cement paste.

= Cement degradation by expansive products formation =

Hardened cement paste contains several hydrated phases whose volumetric mass are lower than that of the corresponding mother anhydrous phase originally present in the clinker. So, some of the neoformed phases have a higher molar volume, meaning they are expansive and swell as a consequence of the hydration process. Two time-dependent situations have to be considered to assess the potential fissuration hazard related to the formation of swelling phases: (1) immediate swelling during cement setting, and (2) delayed crystallisation after setting of the hardened cement paste.

When swelling immediately occurs when the cement is not yet fully set and the cement paste still plastic, the expansive phases can freely swell in the partially-hydrated clinker aqueous suspension and no or no-significant stress develops in the medium.

In contrast to the former favourable case, when the expansion happens at a later phase after the cement paste has set and has lost its original plasticity, physical damages can result from the delayed swelling. The swelling induces then tensile stress in the hardened cement paste and associated strains leading to the formation of micro-cracks and fissures in the hardened cement paste. In the most severe cases, hardened cement paste can totally lose their strength and ultimately fail.

Several reactions in cement produce swelling phases immediately during cement setting or after a more or less long time interval. Two well known detrimental reactions are respectively those leading to the formation of calcium silicate hydrate (CSH) and ettringite (AFm) in cement. The reagents and products respectively involved in these reactions are given in table 1.

Table 1: formation of expansive products capable to cause cracks in hardened cement pastes.

One of the simplest way to avoid cement degradation induced by the neoformed swelling phases, is to prevent their formation by limiting the initial inventory of one of the reactive phases present in the cement (or in the aggregate), or better, by reducing both simultaneously.

= Alkali-silica reaction =

The alkali-silica reaction (ASR), also known as pozzolanic reaction because it is the reaction occurring in ancient Roman mortar made of lime mixed with glassy volcanic ashes, can be schematically written as:

Ca(OH)2 + SiO2 + 2 H2O —> Ca(H2SiO4) · 2 H2O = CSH

Most of the observed cement attack result from uncontrolled modifications in cement chemistry or aggregate mineralogy, or a poor combination of both in the worst cases. Use of cement with high alkali content and/or aggregates rich in amorphous silica favours the formation of delayed swelling CSH in the hardened cement paste. Low alkali cement combined with the absence of silica will prevent any delayed swelling. Intermediate situations are also possible as illustrated in Table 2.

Table 2: four cases to assess the probability of ASR reaction in hardened cement paste.

It must be noticed that a low, but non-negligible, content of amorphous silica can be present in the intergranular “cement” of some silica-rich limestone. It is the case of Kieselkalk and of some carboniferous limestones of the region of Tournais in Belgium. Selecting calcium carbonate as aggregate is thus a necessary but not sufficient choice to reduce the content of amorphous silica in the aggregate. Appropriate specifications of aggregate composition and correct quality controls of limestone aggregates are mandatory to guarantee that they do not contain amorphous silica.

= Cement sulfate attack =

Cement sulfate attack can be abbreviated in the cement chemists notation in the following ways:

3 CaO·Al2O3 + 3 CaSO4 · 2 H2O + 26 H2O —> 6 CaO·Al2O3·3 SO3· 32 H2O

C3A + 3 CSH2 + 26 H —> C6AS3H32

C3A + 3 gypsum + 26 H2O —> ettringite

One notices that ettringite formation consumes 32 water molecules and that 3 mol of gypsum are necessary to combine with one of C3A.

Gypsum is added to control the cement settings by slowing down the C3A hydration rate. If the sulfate addition is to low, flash setting occurs while if it is too high, false setting happens. Gypsum is often added to clinker in the range 2 –4 weight percent, but not higher than 6 wt. %.

The best way to prevent delayed ettringite formation is to use low C3A clinker, requiring thus minimal sulfate addition to control its hydration. The lowest the content of both reagents, the highest the sulfate resistance. In sulfate resistant cement C3A is replaced by C4AF whose slower hydration rate does not required sulfate addition. Table 3 illustrates the combination of different cases that can be encountered for different contents in C3A and SO42–.

Table 3: four cases to assess the probability of delayed ettringite formation in hardened cement paste.

The hereabove-mentioned cases mainly deal with the prevention of internal sulfate attack (ISA). Reducing C3A and SO42– inventories initially present in cement is a very effective way for acquiring sulfate resistance.

However, high sulfate concentration can be present in extreme environments, such as in the excavation disturbed zone (EdZ) affected by pyrite oxidation. In this case, external sulfate attack (ESA) may occur. Alkali (Na, K) sulfates may react with portlandite (Ca(OH)2) in the hardened cement paste, leading to the crystallisation of expansive minerals such as gypsum and thaumasite. Thaumasite, also known as the cement plague, develops along a more complicate reaction pathway involving several cement phases and different steps. To minimise ESA, the best way is likely to limit the sulfate inventory present in the surrounding environment, so, in other words, avoiding pyrite oxidation in the EdZ. This can be achieved by appropriate excavation techniques minimising fissuration around galleries and by closing the galleries as quick as possible after waste emplacement to limit the extent of pyrite oxidation in EdZ. Leaving galleries open for decades or centuries for the sake of long-term waste retrievability in some reversible disposal concepts is not an advised option because of salts accumulation around the gallery lining.

Sulfate attack could led to the formation of cracks in the concrete of the supercontainer, directly exposing the metallic overpack surface to free water filling open fractures in concrete. If thiosulfate and sulfide species would be concomitantly produced by sulfate-reducing bacteria (SRB) in the EdZ around the galleries, they could diffuse without retardation in the free water and rapidly reach the metallic barrier. Carbon steel could be exposed to localised corrosion: pitting induced by S2O32– and stress corrosion cracking (SCC) caused by HS– and S2–. This could jeopardize the integrity of the metallic barrier and in case of its failure make possible water ingress inside the waste package, leading to its accelerated degradation and finally to the release of radionuclides in the surrounding water.

= Recommendation for performance assessment =

The case of external sulfate attack (ESA) should be considered in the reference evolution scenario (RES), and not only the carbonatation process.

= References =

Lea’s cement chemistry (2008) see diagram p. 267.

Taylor (1997) Cement chemistry. see p. 218.