2.1.1. Aggregate
The principal aggregate used in this study is from silica-limestone alluvial deposits from the area of Washington, D.C., USA. It is rather rounded in shape. The two sizes (0-4 mm sand and 4-10 mm coarse aggregate) have a specific gravity (in the dry state) of 2.61 measured following ASTM C127 [11] and C 128 [12]. The respective water absorptions are 0.6 and 0.7% (measured using ASTM C-127 [11] and C128 [12]), the dry packing density values are 0.715 and 0.612 (measured using the vibrational procedure 1 [13]). To assure continuity in the size distribution of the mixtures, a round fine sand (0.106 to 0.075 mm) from Ottawa, Illinois, USA 2 (density 2.65, water absorption 0%, packing density 0.659) was used. This sand corrected the deficiencies in gradation of the alluvial sand by adding fine particles with diameters between the fine sand fraction and the largest cement particle sizes. The grading curves of the aggregates, obtained by sieving, are given in Figure 2. For the mortars, the aggregates were screened to give a maximum 2.5 mm diameter (Figure 2).
In order to model more precisely the contribution of the sand to the packing density of the mixtures, the packing densities of three granular fractions of the sand were also measured. The values were 0.625 for the coarse fraction (1.25 to 4 mm), 0.621 for the middle fraction (0.315 to 1.25 mm) and 0.581 for the finest fraction (0.125 to 0.315 mm).
The packing density mentioned in the paragraph above is defined as the maximum solid material in a unit volume. This factor is 1 for solid material (no air) and 0 for air only.
Figure 2. Size distribution of the aggregates used in the mortars and concretes
2.1.2. Cement
A Type I/II portland cement (ASTM C150 [14]) from Lehigh Portland Cement Co. (Allentown, PA)2 was used for all tests. Its specific gravity was measured to be 3.15. Its compressive strength was 62 MPa at 28 days (following the ISO mortar test [15]). The chemical composition is given in Table 1 and the size distribution (measured in alcohol using a laser granulometer) is given in Figure 3. The packing density, calculated from the water demand test described in [13], is 0.565 without admixture and 0.606 in the presence of 1% HRWRA (total solid by mass of cement).
2.1.3. Silica fume
Silica fume, obtained from Elken 3, was used in slurry containing 54% solids by mass. The specific gravity of the solid was assumed to be 2.2. The size distribution measured using the sediograph method in the presence of lime water and 4% HRWRA shows the expected high degree of fineness of the silica fume (which was 77% by mass finer than 0.25 µm, Figure 3). The chemical composition of the silica fume is given in Table 1.
|
REFERENCE |
Cement |
Silica Fume |
|
Silica |
SiO2 |
21.29 |
96.68 |
|
Aluminum Oxide |
Al2O3 |
4.42 |
0.45 |
|
Titanium Oxide |
TiO2 |
0.22 |
-- |
|
Ferric Oxide |
Fe2O3 |
2.87 |
0.22 |
|
Calcium Oxide |
CaO |
63.83 |
1.24 |
|
Magnesium Oxide |
MgO |
1.78 |
0.24 |
|
Sodium Oxide |
Na2O |
0.21 |
0.24 |
|
Potassium Oxide |
K2O |
0.76 |
0.49 |
|
Sulfuric trioxide |
SO3 |
3.03 |
-- |
|
Chlorine |
Cl |
-- |
-- |
|
Sulfur |
S |
0.07 |
-- |
|
Insoluble Residue |
RI |
0.20 |
-- |
|
Loss on Ignition at 1000ºC |
PAF |
0.77 |
-- |
|
Manganese oxide |
MnO |
0.05 |
0.04 |
|
Total |
99.50 |
99.60 |
Figure 3. Size distribution of the binders used in the mortars and concretes
2.1.4. HRWRA
A sulfonated naphthalene in an aqueous solution with 40% dry extract by mass (commercial name is DARACEM 19 made by the W.R. Grace and Co.) was used as a HRWRA in half of the mixtures. For the rest of the mixtures, no HRWRA was used.
2.2. Methodology
2.2.1. Preparation of the concretes
The concretes were made using a vertical-axis pan mixer (manufactured by Lancaster International, model LWD 4) with a maximum capacity of 16 liters. The aggregates were oven dried before use. All of the dry materials were introduced into the pan-mixer and mixed dry for one minute, then the following mixing schedule was used:
Ordinary concretes:
Concretes with HRWRA:
The total mixing time was 3.5 minutes for the concretes without HRWRA, and 5 minutes for the concretes with HRWRA.
For the concretes with silica fume, the silica fume slurry was added to the dry materials before mixing was started, except for the mixture with the highest dose of silica fume (FS 30), for which the slurry was previously mixed with water containing the first dose of HRWRA.
2.2.2. Using the BTRHEOM rheometer
The experimental plan was designed so as to make a reasonable survey of the full range of workable mixtures of the chosen components. It was expected that segregation problems would be encountered for some compositions. For this reason, the following measurement procedure was selected. After the rheometer chamber was filled, the material was vibrated for 15 seconds at a frequency of 40 Hz. Then, rotation of the blades was begun and increased to a maximum velocity of 0.8 rps (revolutions per second). The relative velocity was reduced in four stages distributed nearly uniformly over the interval from 0.8 to 0.2 rps. Five torque measurements were made at each velocity, and the rheological characteristics of the concrete without vibration were determined. The whole test lasted about 3 to 4 minutes. Because of the abnormal characteristics of a certain number of mixtes, it was decided not to perform rheometer test under vibration since interpretation of the results would have been impossible. At the end of each rheometer test, the appearance of the top horizontal surface of the concrete was noted. This was done by assigning a grade from 0 to 3 based on a qualitative assessment of the amount of bleeding and the segregation caused by rise of the gravel (0 = no bleeding, 3 = severe bleeding).
The BTRHEOM is sealed by a fabric ring seal. This seal creates friction that increases the measured torque. Therefore, before each test of a given mixture, calibartion was performed using water to measure the contribution of the seal to the overall torque measurement. The concrete test was then corrected automatically by the software (ADRHEO) provided with the rheometer using the protocol set forth in Ref. [5].
The experimental plan consisted of systematically characterizing a reasonable range of the mixtures that could be made from the three basic materials: coarse aggregates, combined sand (a mixture of alluvial sand and fine silica sand in fixed proportions) and cement. Based on a certain number of dry compositions (considering only the volumetric fractions of soild materials), three wet mixtures were made by varying the water content so as to cover the range of consistencies that could be characterized by the rheometer. The lower limit was found to correspond to a slump of about 100 mm for mixtures without HRWRA [3], while the upper limit was reached when bleeding was considered to be excessive. A few mixes were also done using silica fume as described in section 2.3.4.
In this series, both the coarse aggregates and the two sands were used, with the exception of the HRWRA. The optimum mixture (called central, hereafter) was calculated using software (Rene-LCPC version 5) developed at Laboratoire des Ponts et Chaussees (LCPC) [13]. The composition of the central mixture was designed to obtain the maximum dry packing density, but with a slight excess of cement in order to minimize the bleeding in all of the mixtures under-dosed with sand. It was the optimization of the packing density at a fixed cement content that led to adopting the proportion of fine sand to alluvial sand of 30% by mass, a value that was maintained throughout the series. The other dry mixtures were generated on the basis of the central mixture by changing one or both of the following two parameters: the volumetric proportion of cement and the volumetric ratio of sand to total aggregate 5 (see Figure 4). It should be noted that the latter ratio was 100% for three dry combinations that were mortars. For these mixtures, to avoid the bleeding that was produced with a cement dosage equivalent to those for the concretes, the proportion of cement by volume was increased as shown in Figure 4. Finally, the combination corresponding to the lowest dosage of cement and the highest dosage of gravel was not made because of the segregation problems, i.e., sedimentation and bleeding, that would have unavoidably occurred.
In order to determine the dosages of water to produce the desired slumps, we began by adopting a dosage corresponding to the porosity of the dry system (provided by the Rene-LCPC software) plus a fixed additional amount of water. Based on the slump obtained, the water dosage was then reduced or increased, the objective being to obtain three concretes (or mortars) of different consistencies for each dry combination. This procedure has the advantage of conserving materials. On the other hand, we sometimes reached the operating limits of the rheometer. However, knowledge of the limits of the applicability of the methodology was also an objective of this study. The composition of mortar and concrete mixtures obtained are given in Appendix I. In addition to the dry compositions (given in mass percentages), the proportions by mass per cubic meter, based on the assumption that the entrapped air content was 1%, are also presents. It should be noted that two additional concretes were produced for the central mixture, bringing to a total of five the number of concretes having the optimum dry composition but variable water content. In Appendix I, the numbers in the mixture identification names correspond to the numbers shown in Figure 4.
Figure 4. Experimental plan. Method of calculating the proportions of dry materials in the series of ordinary concretes and concretes with HRWRAs. The cement content of the mortars (compositions 6-10-11) have been increased to reduce bleeding. S* is the volumetric ratio of sand to total aggregate and C* is the volumetric proportion of cement of the optimum or central mixture.
2.3.2. Concretes with admixtures
In this category of concretes, dry mixtures were generated on the basis of the same mixture design principle as used for mixtures without HRWRA. However, an attempt was made to obtain a higher range of slumps, as is the case in industrial practice with concretes containing HRWRA. To all of these mixtures, 1% of dry extract of HRWRA based on mass of cement was added. This amount, chosen on the basis of results from a series of "mini-slump" [18] type tests on cement paste (Figure 5), is probably close to the saturation dosage based on the AFREM grout method [19], developed at LCPC. The saturation dosage could be interpreted as the maximum dosage of HRWRA beyond which no further increase in the mini-slump value is observed. This dosage of HRWRA ensured against the rapid loss of workability, which would have jeopardized the tests, as well as assuring maximum deflocculation of the cement particles in all of the mixtures, regardless of their composition. However, this high dosage of HRWRA frequently results in excessive bleeding. Moreover, for the lean compositions (low cement content), such as those of groups BHP2 and BHP3, it was not possible to obtain a high slump. Apparently, the water that was in excess of the porosity of the dry mixture, bled instead of separting the particles and "lubricating" the mixture. The compositions of the mortars and concretes of the series with HRWRA are shown in Appendix II.
2.3.3. Intermediate concretes
Although the experimental program was mainly devoted either to concretes without HRWRA or with high dosages of HRWRA, it was also important to examine the effects of intermediate HRWRA dosages, as least, for a limited number of mixtures. Most modern commercial concretes are found in this range. Therefore, four mixtures were generated by simple linear interpolation between the central mixture without HRWRA and its corresponding mixture with HRWRA. Thus, the concrete called "80% BO" corresponds to a mixture (by mass) composed of 80% of the central mixture without HRWRA and 20% of the central mixture with HRWRA (Appendix III).
The silica-fume concretes are another example of mixtures that are of practical interest but are difficult for the rheologist to deal with, because the silica fume particles are for the most part finer than one micrometer (Figure 3), and so are colloidal. At this scale, surface forces can play an important role, making their behavior more complex. Therefore, from the central mixture with HRWRA, we generated four additional concrete mixtures with dosages varying from 7.5% silica fume by mass of the cement to a maximum of 30%, by incremental steps of 7.5%. Obviously, the highest dosage is clearly excessive (too expensive and excessive changes in other properties) for most practical uses of high-performance concrete. The purpose was simply to scan a large range of silica fume dosages and to obtain mixtures in which the silica fume would contribure significantly to the rheology of the system. The dosage of HRWRA was increased at the rate of 4% of the mass of silica fume to maintain a mixture saturated with HRWRA. The compositions of the silica-fume concretes are shown in Appendix III. The quantities of silica fume given in Appendix III are those of the slurry, which is 54% solids by mass.