Subject matter

Objectives:

• Assessing the difference in CH and C–S–H carbonation proportioning and internal H2O production between different accelerated carbonation conditions and natural carbonation for concrete with/without SCMs, and including alkali-activated binders as an extreme case of high-volume SCM addition.
• Determining the changes in mineralogy, evolution in microstructure and transport properties as a function of exposure time and carbonation degree under accelerated test conditions for concrete with/without SCMs.
• Quantifying the effect of mechanical loads on the carbonation resistance of SCM-containing concretes through the development of standardized test setups for this combined attack mechanism.
• Optimizing existing models for carbonation-induced steel depassivation in blended and alkali-activated cement concretes to make them account for all factors mentioned in the previous steps to achieve a more accurate service life prediction.

Material: cement paste, mortar, concrete with supplementary cementitious materials (fly ash, slag, silica fume, metakaolin, natural pozzolans, limestone powder, special SCMs, ternary blends), alkali-activated materials based on these SCMs as precursors

Specific aspects of interest: material properties, mineralogy, microstructure, transport properties, durability, service life, effect of test parameters and preconditioning

Level of investigation: literature review, experimental studies, service life models

Terms of reference

• The TC is supposed to run for 5 years
• Members will be recruited from academia and industry, based on their experience with SCMs in concrete and/or carbonation studies
• The work will include literature research, exchange of good practices information, recommendations, journal publications and/or state-of-the-art report, test programs and inter-laboratory comparisons
• The TC work will be relevant to industry since they often want to assess durability of their proposed concrete compositions based on accelerated laboratory experiments. Hence it should be sure that acceleration does not change the material’s reaction to the environmental conditions. Carbonation of concrete with SCMs, and alkali-activated concrete, is an important consideration for practitioners.

Detailed working programme

Effect of SCMs on carbonation

Supplementary cementitious materials (SCMs) have already a long history of being used as partial replacement of ordinary Portland cement (OPC) in concrete. When focusing on the durability of high-volume SCM concrete, special attention should go to environments subject to atmospheric CO2 ingress and carbonation-induced steel corrosion. In OPC concrete, CO2 dissolves in the pore solution to form carbonic acid which reacts with Ca(OH)2 (CH) and hydrated silicates in the cement paste, forming mainly CaCO3. Although the latter results in a more dense microstructure, the carbonation reaction also implies a significant drop in pore fluid alkalinity which can disrupt the protective passivation layer on embedded steel and cause active corrosion. In SCM concrete, pozzolanic reactions also consume CH. As a result, less CH is available and the carbonation front moves inwards faster. Moreover, the lower CH availability will result in more C–S–H carbonation. In contrast with CH carbonation, C–S–H carbonation does not densify the pore structure. Decalcification and polymerization rather coarsen it. Also note that CH carbonation precipitates mainly well-crystallized calcite, while its amorphous and metastable polymorphs (vaterite and aragonite) are more likely the result of C–S–H carbonation. SCMs result in secondary CSH of lower Ca/Si that also incorporates more alkalis. With alumina-bearing SCMs, C-A-S-H is formed. These changes in composition can also affect carbonation. Alkali-activated concretes, which in general do not contain CH, will progress directly to carbonation of the key strength-giving binder products (C-A-S-H or N-A-S-H type gels) once the carbonic acid has formed in the pore solution. Recent work has shown the importance of layered double hydroxide phases (AFm and/or hydrotalcite groups) in controlling carbonation rates in blended and alkali-activated cements, but this needs further systematic assessment in both cases.

Also, the method for determination of carbonation depth and the degree of carbonation can be discussed (studied with TGA or titration methods) relative to the pore solution chemistry of OPC, blended and alkali-activated binders. Note that phenolphthalein will potentially be banned due to its identified carcinogenic properties.

Effects of accelerated tests at high CO₂ concentration

Furthermore, the higher susceptibility of SCM binders to carbonation is usually concluded from accelerated tests at high CO2 levels. The applied CO2 concentrations in such tests range from 1% to 100%. This is a much broader and higher value range than measured in situ. Relative humidity (RH) and temperature during tests also strongly differ with the varying meteorological conditions in reality.

For OPC binder systems it has been investigated whether accelerated test conditions could alter the carbonation process. Polymerization of C–S–H after carbonation increases with an increasing CO2 concentration. When carbonating at 0.03-3% CO2 there is remaining C–S–H gel. When carbonating at 10-100% CO2 the C–S–H gel completely disappears. The presence of unhydrated phases and ettringite is also significantly affected from a CO2 concentration of 3% onwards. Therefore, various researchers claim that the maximum allowable CO2 concentration for accelerated carbonation should be 3%. Nevertheless, these observations only hold true for OPC pastes and their validity should also be checked for SCM binders. For alkali-activated binders, an upper limit of 1% has been indicated, but ongoing work in TC 247-DTA has shown that 1% and 4% CO₂ do give comparable ranking of different concretes in a round-robin test. Attention needs to be paid to both the carbonation of CH and C–S–H, and to their corresponding effects on the microstructure (densification versus coarsening). Performing carbonation tests at elevated CO2 concentrations may overestimate the C–S–H carbonation, the coarsening of the microstructure and thus the measured carbonation depths and rates resulting from it. Also, the concentration of CO2 may affect carbonation shrinkage. Furthermore, the reaction of the SCMs will provide microstructure densification and lower diffusivity. This would slow down CO₂ ingress, helping to counter the reduced buffering capacity due to the lower CH content. However, this effect could become lost, when using accelerated CO₂ tests. The age of starting the accelerated test is also of importance, especially for slow reacting SCMs.

Additionally, another phenomenon should also be accounted for. The production of CaCO3 always coincides with the release of water. When carbonating concrete at a high CO2 level, the amount of water produced could be more than the porous matrix is capable of expelling in the same time interval, which would slow down further CO₂ ingress.

Scrivener and co-workers (proceedings of DBMC 2017) furthermore showed that the application of a certain relative humidity during accelerated carbonation tests has an effect on the microstructure as well, and Bernal et al. (Mater Struct 2015) showed likewise for alkali-activated concretes. In the work of Scrivener and co-workers, relative humidity of 70% induced a coarsening of the pore structure, and carbonation itself a densification, both for pastes with/without SCMs.

Determination of the prevailing mechanism thus remains a key task and the first efforts of the TC will focus on the effect of carbonation conditions for cementitious material with and without SCMs on the microstructure and phase composition.

The TC activities would include a round-robin test (in laboratory and with concretes naturally exposed to open air environment at different exposure sites/countries). This should be of great interest since EN 206 connects XC exposure classes only with the strength grade of the concrete. Furthermore the environmental impact (both humidity and temperature) is taken into consideration in the probabilistic design approach of fib MC2010 but not in EN206.

Modelling of carbonation

Prediction models for carbonation-induced steel depassivation can range from the very basic square-root-time law derived from Fick’s first law of diffusion, over more accurate physical models considering the microstructural features and the carbonation of the different hydrated phases, to probabilistic engineering models which take into account both concrete and environmental properties. The features found in the previous tasks will be included in the relevant models. The TC can formulate the need for adapted models for concrete containing SCMs, and a benchmarking exercise can be started to compare models with data produced in the committee. The two steps of transport and reaction can be included separately, including consideration of alkali migration in alkali-activated concrete which may also influence the carbonation progress. 

Effects of combined actions: load + carbonation

In standard carbonation tests, microstructural effects of mechanical loads are not accounted for. The work of TC-TDC has shown that the pore structure can be seriously modified under the influence of mechanical loads and that this effects chloride diffusion. Similarly, the carbonation rate will depend on the loading conditions. In presence of compressive stresses not exceeding 50% of the ultimate load level, there should be closure of small microcracks and a reduction in carbonation rate. Higher compressive loads cause new cracks which increase the carbonation rate. In the tensile loading mode, moderate tensile stresses already induce new microcracks which compromise the carbonation resistance. Although the phenomenon is quite well understood, there is still a lack of standardized test methods in that perspective. Hence the TC will also build further upon the expertise gathered within TC-TDC (mainly regarding the combination of load and chloride ingress) and continue its work for carbonation under load.