Durability of concrete and cement composites

The ‘Thaumasite Expert Group Report’ and its ramifications

Because of extensive thaumasite sulfate attack found in Gloucestershire, UK during the late 1990s, particularly on bridges on the M5 motorway, a `Thaumasite Expert Group (DETR)’ was set up by the UK government to examine the situation. The Group reported its findings in 1999 (Department of Environment Transport and the Regions, 1999), and this was followed by some annual reviews. The Report covered risks, diagnosis, remedial works and guidance on new construction. Reassurance was offered that the number of structures in the UK potentially at risk of thaumasite sulfate attack (TSA) was relatively small.

Importantly, TSA was distinguished from mere thaumasite formation (TF). Small occurrences of thaumasite are widespread and do not pose any structural risk in the overwhelming majority of instances. Nevertheless, there are occa- sions when concrete may be apparently affected by TF upon initial inspection but the affected zones may increase in volume over the ensuing months, in which case the onset of TSA is suspected. In such instances, tests and remedial action may need to be undertaken.

The Thaumasite Expert Group (DETR) Report gave guidance for both below- ground and above-ground construction and relied considerably upon technical investigations carried out at the Building Research Establishment in the UK (Crammond, 2003).
The primary risk factors identified were:

· Presence of a sulfate source including sulfide that might oxidise to sulfate.
· Presence of mobile water (groundwater in the instance of buried concrete).
· Presence of carbonate (generally found in the aggregate).
· Low temperatures, generally below 15ëC.

For new below-ground construction, the report recommended revision of key guidance documents for buried concrete and revisions to British Standards on concrete and aggregates. It also examined the issue of acceptable carbonate levels as either aggregate or cement filler permitted within a buried concrete. Three ranges of aggregates were defined in terms of sources that could generate TSA in Portland cement concretes exposed to moderate levels of sulfate. The Ranges, designated A, B and C, refer to those with sufficient carbonate to generate TSA, low carbonate and very low carbonate respectively.

The Report dealt with the use of Portland limestone cement that can contain 6±35% by weight of limestone filler, recommending that these cements should not be used in conditions where the sulfate concentration of the groundwater (as SO4) was above 0.4 g/l. Also addressed was the fact that the European cement standard BS EN 197-1 (2000) permits up to 5% limestone filler to be incorporated in all `common’ cements (as defined). At the time the Report was issued, the addition of 5% limestone filler had been found to make sulfate resistance marginally worse at cold temperatures, when compared with the behaviour of plain Portland cements. The Thaumasite Expert Group reported that there was insufficient evidence to suggest that this would significantly affect the performance of concretes or mortars in the field.

For above-ground construction, TSA had been identified in the UK in lime- gypsum plasters, sulfate-bearing brickwork, ground-floor slabs and concretes contaminated with sulfates and sulfides. All of these types of deterioration can be caused by failure to carry out recommended good practice, either by using unsuitable materials for given service conditions, by poor workmanship on site, or by failure to keep the structures dry. The Group decided that there were already sufficient recommendations in place to combat TSA in above-ground construction. In particular, common sulfate-bearing bricks should not be utilised in exposed conditions, even if they are subsequently rendered, because of poor water-tightness through poor detailing or defective rendering.

TSA may develop as long as the pore solution at the reaction front is kept cold and is maintained at a pH of above 10.5. However, sometimes the deterioration, once started, can continue even if the pH drops below 10.5. This appears to be due to the situation that can arise when thaumasite becomes unstable at around pH 7 and degrades to give secondary calcite (`popcorn calcite’) of irregularly shaped spheres as the ultimate deterioration product of TSA (Crammond, 2003).

Work carried out at the University of Sheffield showed that thaumasite can readily form at low temperatures, not just with Portland limestone cements, but also with limestone filler (up to 5% by weight) in Portland cements (Hartshorn et al., 1999). These results suggested that Portland limestone cements should not be utilised where there is a high risk of TSA taking place, although they have been employed successfully over many years in France, for example, where their use is particularly recommended for internal walls of buildings. In such an application, however, the temperatures are unlikely to fall below ca. 15ëC over lengthy periods and TSA is therefore unlikely to be a problem with normal sound workmanship.

The three-year review of the Thaumasite Expert Group Report (Nixon and Longworth, 2003) noted that new guidance documents (Building Research Establishment, 2001; British Cement Association, 2001; BS 8500-1 and -2, 2002) have increased awareness of TSA and how to achieve preventative measures and drew attention to various points including the following:

· Two new field cases of TSA had arisen (in the UK) in buried concrete containing siliceous aggregates. These justify concerns expressed in the Report that concretes whose aggregates contain little or no carbonate can be affected by TSA if an external source of carbonate ions is available, e.g. from groundwater where dissolved bicarbonate is continually available.
· Cements with a substantial proportion of ground granulated blast furnace slag (ggbs) had so far been shown to offer good resistance to TSA.

In the 2005 revision of BRE Special Digest 1, it is recognised that carbonate required for TSA can be supplied externally, principally from bicarbonates in groundwaters (see equations 8±10, Section 4.6.2), and so the concrete quality is  not relaxed for carbonate-free aggregates, as in the earlier version of this guidance.

In some comments on the aforementioned three-year review, the question of pfa in relation to ggbs for minimising TSA occurrences has been mentioned (BS 8500-2, 2002). Ggbs exhibits delayed hydraulicity, being protected by surface films from significant reaction at the beginning of hydration, whilst pfa is not intrinsically hydraulic and needs a source of OHÿ ions to instigate pozzolanic reactivity in the presence of Ca2+ ions. Pfa may not always be effective in com- bating TSA if the pozzolanic reaction has not already proceeded to a significant extent to form additional C-S-H as at ca. 5ëC. In such an event the pfa would essentially act as a filler, for example if the OH- ions are depleted by reaction with CO2 to form soluble hydrogen carbonate ions: OHÿ + CO2!HCO3ÿ rather than initiating more C-S-H formation. A Portland-based cement with a high proportion of unreacted pfa would be likely to be vulnerable to TSA in such circumstances, since the permeability would not be satisfactorily reduced. However, formation of thaumasite is slow and, when (as in most cases) the pozzolanic reaction is well under way by the time thaumasite is likely to be produced, then the pfa may perform a similar function to ggbs as an extended cement component. More reactive pozzolans such as microsilica and metakaolin do not appear to be affected as pfa can at times be. Initiation of extended hydration by slags and pozzolans (like pfa) needs to be more fully investigated in the context of potential for TSA and its minimisation. It should then be possible to achieve optimal TSA resistance in practice under given field conditions (Bensted, 2003b). An interesting study (Sims and Huntley, 2004) has shown that TF or TSA can occur even when the primary risk factors (Department of Environment, Transport and the Regions, 1999) are not all obviously present. A source of external or internal sulfate was always present in these cases, but water was not always abundant or mobile. There was not always a direct internal source of carbonate and temperatures were not always low. Results from an extensive programme of laboratory tests on concrete cubes of water/binder ratio 0.45, made with three different cements and moist cured for 28 days before being exposed to various sulfate solutions at 5ëC, have also been published recently (Zhou et al., 2006). Some of the main conclusions of this work are summarised below:

· For cubes stored at pH 12 in solutions of calcium sulfate (1.4 g SO4/l) + magnesium sulfate (1.6 g SO4/l) ± corresponding to Design Sulfate Class DS- 3 (Building Research Establishment, 2005) ± TSA was observed within 5 months and significant damage within 12 months.
· The attack was found to be severe for cubes made with Portland limestone cement containing 20% by weight of limestone filler, but also evident in cubes from OPC (with 5% by weight of limestone filler), and SRPC with neither added limestone filler nor containing limestone aggregate. It was

therefore suggested that both SRPC and PC containing 5% by weight of limestone filler should be used cautiously in buried concrete liable to come into contact with sulfate-containing groundwater.
· The presence of acid did not promote formation of thaumasite. Although thaumasite was observed in lesser amounts in cubes immersed in sulfuric acid solution at 5ëC, the nature of the corrosive attack was different from TSA. It was suggested, however, that initial attack of concrete by acid could result in its being highly vulnerable to TSA, should this be followed by alkaline high  sulfate conditions.
· Both C-S-H gel and calcium hydroxide were consumed during formation of thaumasite, which was often accompanied by formation of gypsum and, when magnesium was present, by magnesium hydroxide (brucite).
· It is likely that the TSA observed in the buried concrete associated with the M5 motorway in the west of England was due to the increased sulfate content from oxidation of pyrite (iron disulfide FeS2) rather than due to low pH associated with the production of sulfuric acid.

In conclusion it appears that, while the state-of-the-art is still not yet clearly defined in all important respects, a great deal of new knowledge has been gained through recent research on thaumasite formation and thaumasite sulfate attack. When fully digested, this should help to minimise future occurrences of this relatively uncommon but insidious form of sulfate attack.


Related Articles