Iron was corroded in anoxic cementitious (pH 13.2) media (filtered porewater or cement slurry) at 80 degrees C; corrosion products were investigated by microscopic techniques. A thin film of magnetite (Fe3O4) formed during the first immersion days. For longer immersion times (90 days) in cement slurry, hydroandradite (HA) (Ca-3(Fe3+)(2)(SiO4)(y)(OH)(4(3-y))) formed a dense outer layer and an heterogeneous inner layer. A thin oxide layer is always observed in contact with metal and is believed to correspond to a protective barrier. Our results suggest that in the conditions of the Belgian supercontainer for radioactive waste disposal, a passive layer can form slowing down corrosion.

The influence of atmospheric carbonation on gas diffusion was investigated using four hardened cement pastes (CEM I, CEM III/A, CEM V/A, and a low-alkalinity binder) with common water-to-binder ratio (0.4). The diffusivity of the non‑carbonated and carbonated pastes with respect to helium and nitrogen was measured at different relative humidities. Carbonation decreased the diffusivity of the CEM I paste, whereas that of the other binders significantly increased after carbonation. These results show the competition between porosity clogging and cracking induced by carbonation. The consequences of carbonation are therefore believed to depend on the considered binder. Clogging dominates in ordinary Portland cement (OPC), leading to a decrease in its transport properties after carbonation; cracking dominates in blended cements, leading to a significant increase in its transport properties after carbonation.

The consequences of accelerated carbonation at 3% CO2 were compared with those of natural carbonation (0.04%). Cement pastes (CEM I and CEM V/A) as well as the three major constitutive phases (C-S-H of different C/S ratios, portlandite and ettringite) were used and changes in the mineralogy, microstructure, water retention and cracking were investigated. The main conclusion was that accelerated carbonation at 3% CO2 was representative of natural carbonation although it promoted the precipitation of metastable calcium carbonate (aragonite and vaterite) in place of calcite. The results also showed that the presence of aragonite and vaterite were characteristic of the carbonation of ettringite and C-S-H respectively.

The impact of temperature on carbonation was investigated in laboratory conditions using a device developed for this purpose. Two hardened cement pastes (CEM I and CEM V/A) were tested between 20 degrees C and 80 degrees C at different levels of relative humidity (RH). The carbonation rate of the CEM I increased with temperature, whereas that of CEM V/A reached a maximum at around 50 degrees C.

16 In unsaturated conditions, the durability of concrete structures is strongly dependent on the evolution 17 of the amount of free water within concrete porosity. Reliable durability assessment of concrete 18 structures in relation to their environment thus requires accurate unsaturated water transport 19 description as well as reliable input data. The effect of carbonation on water transport remains poorly 20 studied and data are lacking. It was then the purpose of this article to acquire all the data needed to 21 describe unsaturated water transport in carbonated cementitious materials (porosity, water retention 22 and unsaturated permeability). Four hardened pastes made with four different binders were 23 carbonated at 3% CO2 to ensure representativeness with natural carbonation. Beyond the modification 24 of the water retention curve and porosity clogging, significant microcracking due to carbonation 25 shrinkage was observed. The consequence on permeability highlighted a competition between 26 porosity clogging and microcracking that was dependent on the initial mineralogical composition. 27

Calcium sulfoaluminate (CSA) cements are potential candidates for the conditioning of radioactive wastes with high sodium borate concentrations. This work thus investigates early age hydration of two CSA cements with different gypsum contents (0 to 20%) as a function of the mixing solution composition (borate and NaOH concentrations). Gypsum plays a key role in controlling the reactivity of cement. When the mixing solution is pure water, increasing the gypsum concentration accelerates cement hydration. However, the reverse is observed when the mixing solution contains sodium borate. Until gypsum exhaustion, the pore solution pH remains constant at ~ 10.8, and a poorly crystallized borate compound (ulexite) precipitates. A correlation is established between this transient precipitation and the hydration delay. Decreasing the gypsum content in the binder, or increasing the sodium content in the mixing solution, are two ways of reducing the stability of ulexite, thus decreasing the hydration delay.

In unsaturated conditions, the durability of concrete structures is strongly dependent on the evolution of the amount of free water within concrete porosity. Reliable durability assessment of concrete structures in relation to their environment thus requires accurate unsaturated water transport description as well as reliable input data. The effect of carbonation on water transport remains poorly studied and data are lacking. It was then the purpose of this article to acquire all the data needed to describe unsaturated water transport in carbonated cementitious materials (porosity, water retention and unsaturated permeability). Four hardened pastes made with four different binders were carbonated at 3% CO2 to ensure representativeness with natural carbonation. Beyond the modification of the water retention curve and porosity clogging, significant microcracking due to carbonation shrinkage was observed. The consequence on permeability highlighted a competition between porosity clogging and microcracking that was dependent on the initial mineralogical composition.

We develop and apply in this study a chemo-transport-mechanical model for simulating the external sulfate attacks in Portland (CEM I) cement pastes and mortars. Basically, this degradation consists in the simultaneous decalcification of the hydrated phases resulting from leaching processes, and the migration of sulfate ions within the material and its subsequent interactions with these phases. The sulfate uptake leads generally to ettringite precipitation mainly from monosulfate, which in turn may produce intense macroscopic expansions and cracking. In our approach, crystallization pressures arising from the restrained growth of monosulfate crystals due to the confinement of the surrounding C–S–H matrix are assumed to initiate the observed macroscopic expansions. A macroscopic strain tensor evaluated from the volume fraction of supplementary precipitated ettringite is further introduced in the mechanical behavior law for explicitly reproducing the macroscopic expansions. Analytical homogenization schemes are applied to estimate both mechanical and diffusive properties from the local volume fraction of solid phases. The numerical platform Alliances is then used for solving both reactive transport and mechanical coupled problems, and is applied to the simulation of laboratory tests consisting in prismatic mortar specimens immersed in solutions containing sodium sulfate and subjected to free expansions. Comparison of the numerical results with experimental ones in terms of phase assemblage profiles, evolutions of mass changes and expansions shows a correct agreement. Finally, the extension of the model towards cases of restrained displacement conditions is discussed and some modifications regarding the kinetics of ettringite precipitation are proposed for such situations.

A specific cell was designed to monitor simultaneously the evolution of the viscoelastic properties, electrical conductivity, and temperature of a cement paste with ongoing hydration. Hydration of calcium sulfoaluminate cement by demineralised water or by a borated solution was then investigated as an example. Borate anions acted as set retarders but to a smaller extent than with ordinary Portland cement. The delay in cement hydration resulted from the precipitation of an amorphous or poorly crystallized calcium borate, which also caused a rapid stiffening (and thus a loss of workability) of the paste after mixing. The gypsum content of the CSA cement was shown to play a key role in the control of the cement reactivity.

This study proposes two models for accelerated leaching in ammonium nitrate under variable temperature. The first approach assumes a thermoactivated linear relationship between degradation and the square root of time. The second approach introduces, within a phenomenological model, the thermoactivation of diffusion and hydrates solubility. These two approaches enable one to analyse accelerated leaching tests to identify the concrete variability in order to get rid of the influence of temperature. Cette étude consiste à proposer deux modélisations de la lixiviation accélérée au nitrate d’ammonium sous température variable. La première approche suppose une relation linéaire thermoactivée entre la dégradation et la racine carrée du temps. La seconde approche introduit dans un modèle phénoménologique la thermoactivation de la diffusion et de la solubilité rétrograde des hydrates. Ces deux approches permettent de dépouiller des essais de type DANA afin d’identifier la variabilité du béton en s’affranchissant de l’influence de la température.

We have focused on the test conditions influence on accelerated degradation of cementitious materials using ammonium nitrate. PH-buffering and renewal of the leaching solution were studied. PH-buffering appeared not to be very important when the renewal pH remains under eight. Renewal appeared to be the most influential feature. Its absence leads to calcium accumulation in the leaching solution inducing aggressiveness fall. Degradation is then less marked in terms of depth, flux and mineralogy. The resulting porosity increase is also smaller.

The potential of calcium sulfoaluminate (CSA) cement was investigated to solidify and stabilize wastes containing large amounts of soluble zinc chloride (a strong inhibitor of Portland cement hydration). Hydration of pastes and mortars prepared with a 0.5 mol/L ZnCl(2) mixing solution was characterized over one year as a function of the gypsum content of the binder and the thermal history of the material. Blending the CSA clinker with 20% gypsum enabled its rapid hydration, with only very small delay compared with a reference prepared with pure water. It also improved the compressive strength of the hardened material and significantly reduced its expansion under wet curing. Moreover, the hydrates assemblage was less affected by a thermal treatment at early age simulating the temperature rise and fall occurring in a large-volume drum of cemented waste. Fully hydrated materials contained ettringite, amorphous aluminum hydroxide, strätlingite, together with AFm phases (Kuzel's salt associated with monosulfoaluminate or Friedel's salt depending on the gypsum content of the binder), and possibly C-(A)-S-H. Zinc was readily insolubilized and could not be detected in the pore solution extracted from cement pastes.

To assess the potential of calcium sulfoaluminate cement to solidify and stabilize wastes containing high amounts of soluble zinc chloride (a strong inhibitor of Portland cement hydration), a simulated cemented waste form was submitted to leaching by pure water at a fixed pH of 7 for three months, according to a test designed to understand the degradation processes of cement pastes. Leaching was controlled by diffusion. The zinc concentration in the leachates always remained below the detection limit (2 μmol/L), showing the excellent confining properties of the cement matrix. At the end of the experiment, the solid sample exhibited three zones which were accurately characterized: (i) a highly porous and friable surface layer, (ii) a less porous intermediate zone in which several precipitation and dissolution fronts occurred, and (iii) the sound core. Ettringite was a good tracer for degradation. The good retention of zinc by the cement matrix was mainly attributed to the precipitation of a hydrated and well crystallized phase with platelet morphology (which may belong to the layered double hydroxide family) at early age (≤ 1 day), and to chemisorption onto aluminum hydroxide at later age.

The interaction of mortar with sulfate solutions leads to a reaction front within the porous material and to expansion. Thermodynamic modelling coupled with transport codes was used to predict sulfate ingress. Alternatively, “pure” thermodynamic models – without consideration of transport – were used as a fast alternative to coupled models: they are more flexible and allow easy parameter variations but the results relate neither to distance nor to time. Both transport and pure thermodynamic modelling gave comparable results and were able to reproduce the changes observed in experiments. The calculated total volume of the solids did not exceed the initial volume of the paste indicating that not the overall volume restriction leads to the observed expansion but rather the formation of ettringite within the matrix and the development of crystallisation pressure in small pores. The calculations indicate that periodic changing of the Na2SO4 solution results in more intense degradation.

Hydration of a belite calcium sulphoaluminate cement was investigated over one year as a function of its initial gypsum content (variable from 0 to 35%). Particular attention was paid to the influence of the thermal history of the material at early age on its subsequent evolution. Pastes and mortars (w/c 0.55) were either cured at 20°C or submitted for one week to a thermal treatment simulating the temperature rise (up to 85°C) and fall occurring in drums of cemented radwastes. The thermal cycle accelerated the early stages of hydration and mainly decreased the proportion of AFt versus AFm hydrates, especially at low initial gypsum contents (≤20% by weight of cement). It also strongly reduced the compressive strength of gypsum-free specimens (by 35% after one year), and doubled their expansion under water. These results were explained by mineralogical evolutions towards a more stable phase assemblage which included retarded ettringite formation.

Hydration of calcium sulfoaluminate cement at early age was investigated as a function of the gypsum content of the binder, the thermal history of the material, and the ZnCl2 concentration in the mixing solution. Early hydration was strongly accelerated by the presence of gypsum, but lower percentages of reaction were reached after 24h. The slowing down effect induced by ZnCl2, even at a concentration as high as 0.5mol/L, was moderated compared to OPC but had a greater intensity in the absence of gypsum. Unlike what would have been expected for Portland cement, it was shown that the delay of a gypsum-free calcium sulfoaluminate cement resulted from the strong retardation caused by chloride anions, which was partly compensated by the accelerating effect of Zn2+ cations. The mineralogical observations revealed the precipitation of chloro–AFm phases such as Friedel's and Kuzel's salts, but no crystallized zinc-containing phases could be identified by XRD. The thermal history of the samples proved to be a key parameter. Applying a thermal cycle which reproduced the temperature rise and decrease occurring in a massive mortar block accelerated the rate of hydration and mainly modified the proportion of AFt versus AFm hydrates, especially when the binder had a gypsum content below 20%.

ABSTRACT Investigations were carried out to formulate and characterize low-alkalinity and low-heat cements which would be compatible with an underground waste repository environment. Pozzolans (high-silica product such as fly ash and silica fume) and blast furnace slag were added to Portland cement to decrease the cement pore solution pH to 11. The alkali content of the interstitial solution (≈ 1 to 5 mmol/L) of blend was strongly reduced which decreased the pH by more than one unity as compared with control samples. The blends exhibited a low heat output (temperature rise below 20°C). The prepared concrete specimens had a compressive strength higher than 70 MPa after one year of curing in 100% RH, a low shrinkage (−350 to −500μm/m) and it was possible to prepare workable concrete despite the very high amounts of pozzolans in the blends.

ABSTRACT Materials based on Portland cement react chemically with solutions containing sulphates, even at very low concentrations. The originality of the work described in this paper is to have characterised the kinetics of evolution of mineralogy and solution-material chemical exchanges, for a concentration 10.10−3 mol/L of sulphate, representative of underground conditions, which is a deep geological site made up of argillite. Under these chemical conditions which are maintained as constant as possible at the laboratory scale, the main experimental fact established is the precipitation of gypsum in the damaged zone for two cements with very different C3A contents, despite the low sulphate concentration of the aggressive solution. The gypsum forms immediately behind the dissolution front of portlandite, which itself spreads towards the core of the material at a rate of 0.15 mm.d−0.5. The amplitude of the precipitation of secondary ettringite is consistent with the C3A content of the cement. However, such mineralogical changes did not cause macroscopic alteration.