III. Results and discussion
III.1. Correlation between ASTM C1506--09 and DIN 18555--77test methods
The comparison between both methods is shown on Figure 6. Usually, both methods gave rather close results, except for few molecules. Those were three HEC (N1, N2 and N3) that conferred to the mortar a more liquid aspect. The water retention value obtained with the DIN method was close to 90% while the result of ASTM was fairly higher (roughly 97%). This can be explained by a bleeding phenomenon observed for these admixtures during the experiment. Indeed, water seemed to stay above whereas sand is at the bottom.
The ASTM values were generally slightly higher than those obtained with the DIN method.
This phenomenon can be explained by a depression effect. Indeed, the ASTM measurement was performed under vacuum (50 mm of mercury), while the DIN method was a measurement of absorbed water in contact with a filter paper, based on gravity and performed at atmospheric pressure.
All things considered, the DIN 18555-7 and the ASTM C1506-09 methods are two ways to determine water retention that are comparable. In our paper, only the water retention results obtained with the ASTM method are shown.
III.2. Helical geometry: an unusual way to characterize settling mortars
To the extent that it generates a vertical pumping, the helical geometry was used to minimize the sedimentation of particles during the rheological test . Such mixer-type geometry belongs to the category of process geometries that allows the possibility to extract rheological
information directly from torque-rotor speed measurement in batch or semi-batch configuration, during and at the end of the preparation of a complex liquid-like product. Using a Couette analogy, it has been shown that the torque-rotor speed data can be transformed into shear stress–shear rate curves, which are in fairly good agreement with offline measurements obtained in conventional rheometers .
In order to validate the experimental procedure, the results obtained by submitting nonsettling samples to a high velocity gradient before each measurement were compared with results obtained in standard steady state regime i.e. without resuspension steps. These experiments were performed using the vane geometry for a large range of CE-admixed mortars.
For example, for HEMC C3 and C4 (Figure 7), results show that the high shear rate steps did not influence significantly the rheological behaviour of the samples. This comparison also demonstrated that the yield stress noted in the case of HEMC C3 in steady state is not a real yield stress; it is most probably due to sedimentation. These results confirm what is observed with the naked eyes.
Then, to validate the use of the helical geometry, a comparison with the vane was realized for CE-admixed mortars that were not subjected to settling. Figure 8 presents a comparison for two HEMC (C3 and C4) and shows that the rheograms obtained with both geometries are very similar. It confirms that the helical mobile can be used without effect on mortar structure. In consequence, since it allows a better homogenization of the samples, the helical geometry, associated with the resuspension procedure, was used to perform experiments for settling mortars.
 L. Nachbaur, J.C. Mutin, A. Nonat and L. Choplin, Dynamic mode rheology of cement and tricalcium silicate pastes from mixing to setting, Cem. Concr. Res. 31 (2) (2001).