Dust-of-Fracture Aggregate Microfines in Self-Consolidating Concrete

(ProQuest: ... denotes formulae omitted.)

INTRODUCTION

Self-consolidating concrete (SCC) must have sufficient paste volume and proper paste rheology to flow under its own mass, pass through congested reinforcement, and resist segregation.1,2 Paste volumes are usually higher for SCC than for conventionally placed concrete and typically consist of high powder contents and low water-powder ratios (w/p).3 If the powder content of SCC comprises only cementitious materials, the hardened properties may be significantly overdesigned for the application and the materials costs may be prohibitively high. Therefore, this paper explores the use of dust-of-fracture aggregate microfines-defined as material finer than 75 µm produced during the crushing of aggregates- to comprise part of the powder content.

ASTM C33 currently limits the quantity of dust-of-fracture microfines essentially free of clay and shale to 7% of the fine aggregate mass (5% for concrete subject to abrasion). As a result, microfines typically must be washed from manufactured sands, resulting in added costs to wash the sand and to stockpile or dispose of the microfines. Previous research, however, has shown that microfines can be used successfully in concrete at rates up to 25% of the fine aggregate mass.4 In addition, finely ground limestone fillers have been used extensively in SCC to comprise a portion of the powder content.3,5-9 Because their use is restricted by ASTM C33, microfines are usually available for very low cost when compared with cement, supplementary cementitious materials, and finely ground limestone filler.

Although microfines are similar in size to other powder constituents-such as cement, slag, and fly ash-they are often accounted for as part of the fine aggregate volume. When microfines are added to a concrete mixture and accounted for as part of the fine aggregate volume-namely, replacing part of the fine aggregate volume with microfines-the paste volume is increased and the w/p is reduced. The decrease in workability due to the reduction in w/p typically more than offsets any increase in workability due to the increase in paste volume, resulting in a net reduction in workability.10-14 In contrast, when microfines are accounted for as part of the powder volume and the paste volume and w/p are held constant-namely, replacing part of the powder volume with microfines-changes in workability are typically much less and are predominately a function of the particle size distribution, shape characteristics, and clay content of the microfines.15,16 The particle size distribution of microfines is important in the context of how the microfines affect the overall powder particle size distribution.9 The shape characteristics of microfines can vary widely17 and should be considered relative to the other powder materials the microfines are replacing. In one study of granite aggregates from the same source and crushing process, it was found that the shape characteristics of the microfines were similar to those of the coarse aggregates.18 Clays sometimes present in microfines can increase water demand19 and interact with polycarboxylatebased high-range water-reducing admixtures (HRWRAs).20

RESEARCH SIGNIFICANCE

The use of dust-of-fracture aggregate microfines as part of the powder content in SCC can be an economical alternative to using powder comprised only of cementitious materials. Although finely ground limestone fillers have been used extensively in SCC, much less data are available for the use of dust-of-fracture microfines in SCC. This paper presents an experimental evaluation of the effects of dust-of-fracture microfines on the workability and hardened properties of SCC.

EXPERIMENTAL PROCEDURES

Materials and mixture proportions

Six microfines were tested in mortar and concrete. In the mortar study, microfines were used as part of the aggregate volume or powder volume, as shown in Table 1. Accounting for microfines as part of the aggregate volume resulted in constant water-cementitious material ratio (w/cm), reduced w/p, and increased paste volume. Accounting for microfines as part of the powder volume resulted in increased w/cm, constant w/p, and constant paste volume. All mixtures incorporated fly ash at a rate of 25% of the cementitious materials mass.

In the concrete study, a constant total volume of microfines (15% of the fine aggregate volume) was accounted for as either part of the aggregate volume or the powder volume, as shown in Table 2. To compare hardened properties, a second control mixture was evaluated. Control 1 is the baseline case with 35.9% paste volume, a w/p of 1.073 (by volume), and a w/cm of 0.37 (by mass). Control 2 has the same paste volume of 35.9% but a higher w/cm of 0.524 to match that of the mixtures with microfines accounted for as powder volume. All mixtures incorporated fly ash at a rate of 25% of the cementitious materials mass. When microfines were used in concrete as part of the powder volume, they accounted for 29.3% of the powder volume.

The properties of the six microfines are shown in Table 3. Microfines were obtained as settling pond fines from a quarry (LS-02) or were dry-sieved from dry screenings (all other microfines). A limestone manufactured sand was used in all mortar and concrete mixtures and a river gravel coarse aggregate was used in all concrete mixtures (Table 4). The limestone manufactured sand was washed over a No. 200 sieve to remove all microfines. All mixtures included a Type I/II cement, Class F fly ash, and polycarboxylate-based HRWRA. The properties of the cement and fly ash are shown in Table 5. For the microfines, cement, and fly ash, packing density was measured with the drop test21 and particle size distribution was measured with laser diffraction (Fig. 1). The drop test was conducted by placing a known volume of water (0.2 mL [0.0068 oz]) on a dry bed of loosely packed powder. The mass of the resulting agglomerate of water and powder was determined, converted to volumes of water V^sub water^ and powder V^sub powder^, and used to compute packing density f, as shown in Eq. (1)

... (1)

Packing density is primarily affected by the particle size distribution and shape characteristics of the powder. As shown in Fig. 2, the microfines had higher packing densities than the cement but lower packing densities than the fly ash. The addition of microfines to the blend of cementitious materials from the control mixture did not significantly change the combined powder packing densities because the packing densities of the individual microfines were near that of the cementitious materials. The laser diffraction particle size distribution measurements were used to calculate specific surface area (SSA) and span. Because laser diffraction measurements are based on the assumption of spherical particles, the calculated specific surface areas reflect only size and particle size distribution and not shape characteristics. Span was computed as shown in Eq. (2)

... (2)

where d(0.9) is the diameter with 90% passing, d(0.5) is the diameter with 50% passing, and d(0.1) is the diameter with 10% passing. The methylene blue value test was conducted in accordance with AASHTO TP 57 to determine the presence of clays in the microfines. The methylene blue test measures the amount of methylene blue dye adsorbed by clays within a sample of microfines.19

Test procedures

Mortar was mixed in 2.5 L (0.09 ft^sup 3^) batches in a mixer meeting the requirements of ASTM C305. Mortar workability was measured with the mini slump flow and mini V-funnel tests, which are shown in Fig. 3. The dosage of HRWRA was adjusted in each mortar batch to reach a mini slump flow of 230 mm (9 in.).

Concrete was mixed in 70 L (2.5 ft^sup 3^) batches in a rotating drum mixer. Concrete workability was measured with the slump flow test and a vane-type concrete rheometer. The slump flow test was performed in accordance with ASTM C1611 with the cone in the inverted orientation. The dosage of HRWRA was adjusted in each concrete batch to reach a slump flow of 610 to 660 mm (24 to 26 in.). The rheometer22 used a four-bladed vane that was 125 mm (5 in.) in height and diameter. The vane was centered in a 280 mm (11 in.) diameter container filled with concrete to a height of 280 mm (11 in.). The container walls included vertical strips to prevent slippage. The rheometer was used to perform flow curve measurements, which consisted of a 20-second preshear period-where the vane was rotated at 0.60 rps-followed by seven flow curve points measured in descending order from 0.60 to 0.05 rps. Yield stress and plastic viscosity were calculated based on the Bingham model from the seven descending flow curve points.

Compressive strength and static modulus of elasticity were measured on 100 x 200 mm (4 x 8 in.) cylinders in accordance with ASTM C39 and ASTM C469, respectively. Flexural strength was evaluated on 115 x 115 x 395 mm (4.5 x 4.5 x 15.5 in.) simply supported beams with 345 mm (13.5 in.) span length. Third-point loading was applied in accordance with ASTM C78. Rapid chloride permeability was measured in accordance with ASTM C1202 on 50 mm (2 in.) thick specimens cut from 100 x 200 mm (4 x 8 in.) cylinders. Shrinkage was measured in accordance with ASTM C157 on 75 x 75 x 285 mm (3 x 3 x 11.25 in.) specimens with a 250 mm (10 in.) gauge length. After casting, shrinkage specimens were stored at 23 °C (73 °F) and 100% relative humidity for the first 24 hours, stored in limewater solution to an age of 3 days, and then stored at 23 °C and 50% relative humidity for the remainder of the test. Abrasion resistance was measured with the rotating cutter method (ASTM C944) on formed-surface specimens. A double load of 20 kg (44 lb) was applied for 8 minutes to each abrasion specimen.

EXPERIMENTAL RESULTS AND DISCUSSION

Mortar and concrete workability

The effects of microfines on workability were expressed in terms of HRWRA demand and plastic viscosity. For mortar, the mini V-funnel time was assumed to be correlated to plastic viscosity. Decreases in HRWRA demand and plastic viscosity are favorable for SCC workability in the context of the results presented in this paper. Although the plastic viscosity of SCC can be too low, resulting in poor workability, changes in materials and mixture proportions that reduce plastic viscosity are favorable because they can be used to offset other factors that increase plastic viscosity. HRWRA is used to reduce yield stress to ensure self-flow (reflected in slump flow measurements). Therefore, the HRWRA demand for a given slump flow is related to yield stress-namely, the greater the HRWRA demand, the greater the required decrease in yield stress to achieve the low yield stress for self-flow.

For the mortar mixtures, the use of microfines as part of the aggregate volume increased HRWRA demand for all six microfines considered (Fig. 4). The decrease in w/p, offset by the increase in paste volume, was partially responsible for this increase in HRWRA demand. The large difference in performance between microfines at a given microfines content was due to differences in microfines characteristics- including particle size distribution, shape characteristics, and methylene blue value. The mini V-funnel time decreased at 5% microfines contents for all but GR-01 and increased at 20% microfines content for all but LS-06 (Fig. 5). As with HRWRA demand, the changes in mini V-funnel time with increasing microfines content were due to the decreasing w/p, increasing paste volume, and differences in microfines characteristics.

HRWRA demand (expressed in L/m^sup 3^) and mortar V-funnel time were consistently lower when the same total volume of microfines was used as part of the powder volume rather than the aggregate volume (Fig. 6 and 7). This difference was attributable to the fact that the w/p, powder volume, and paste volume were held constant when microfines were used as part of the powder volume. The changes in workability when microfines were used as part of the powder volume were due only to differences in the microfines characteristics. Therefore, in evaluating the effects of microfines, it is important to consider the effects of the microfines themselves separately from changes in mixture proportions.

The use of TR-01 and DL-01 resulted in the largest increases in HRWRA demand; however, the changes in mini V-funnel time were much smaller. In fact, the use of these two microfines as part of the powder volume resulted in slight decreases in mini V-funnel time. The apparent presence of clays, reflected in the high methylene blue values for these two microfines, likely resulted in much of the increase in HRWRA demand. It is known that certain clays can consume polycarboxylate-based HRWRAs.20 Therefore, sufficient HRWRA must be provided to be consumed by the clays and provide dispersion of powder materials. Alternatively, the use of sacrificial agents or certain batching sequences have been reported to mitigate the effects of clay.17,20 Once sufficient HRWRA was provided to offset the apparent effects of the clays, the workability was otherwise acceptable as indicated by the low mini V-funnel times.

In contrast, GR-01 exhibited increased HRWRA demand and mini V-funnel time. Although GR-01 had the lowest methylene blue value, the poor shape and particle size distribution of this aggregate contributed to the poor workability. The poor shape and particle size distribution were partially reflected in the low packing density of the combined GR-01 microfines and cementitious materials (Fig. 2). Powders with high packing density result in less voids space that must be filled with water. In addition, improved shape characteristics not only result in increased packing density, but also less interparticle friction and, thus, improved workability.1 The three limestone microfines performed consistently better than the other three microfines when used as aggregate volume or powder volume. These microfines had low methylene blue values and high combined powder packing densities.

It is important to note that the packing densities of the individual microfines were similar to the packing densities of the combined cementitious materials (Fig. 2). Had the microfines been compared with a cement-only control mixture, it is likely their effect on workability would have been more favorable because of the bigger differences between the packing densities of the microfines and the cement.

Multivariate regression analysis was used to evaluate further the effects of specific microfines characteristics on mortar workability. The models were developed with the data for the microfines used as part of the aggregate volume. The independent variables were a percentage of microfines (pct), packing density (pkg), SSA, span, and methylene blue value (mbv). Full quadratic regression models were developed with a step-wise procedure (p-value of 0.10 for including or excluding terms) for HRWRA demand (R^sup 2^ = 0.98) and mini V-funnel time (R^sup 2^ = 0.96), as shown in Eq. (3) and (4), respectively

HRWRA demand (L/m^sup 3^) = 0.652 + 0.00260(pct)^sup 2^ + 1.289(pkg)^sup 2^ + 0.114(pct)(pkg) - 0.0734(pct)(SSA) + 0.0120(pct)(mbv) (3)

Mini V-funnel time (s) = 6.420 + 2.147(pct) + 0.0140(pct)^sup 2^ - 2.533(pct)(pkg) - 0.0248(pct)(span) - 0.0271(pct)(SSA) - 0.00228(pct)(mbv) (4)

The regression models, which are plotted in Fig. 8 and 9 for 15% microfines, indicate that HRWRA demand was primarily affected by methylene blue value and specific surface area, whereas mini V-funnel time (plastic viscosity) was affected by packing density, methylene blue value, specific surface area, and span. Increasing the specific surface area (higher fineness) significantly decreased HRWRA demand and mini V-funnel time. In evaluating the effects of specific surface area, the combined particle size distribution of all powders should be considered. As shown in Fig. 1, the finer microfines-namely, LS-02 and LS-06- did not overlap the particle size distribution of the cementitious materials to the same extent as the coarser microfines. Additionally, the coarser microfines interacted to a greater extent with aggregates retained on the No. 200 sieve. As a result, the finer microfines had a more favorable effect on combined powder particle size distribution and workability. Likewise, increasing the span resulted in slightly lower mini V-funnel time. Microfines with wider spans had less overlap with the cementitious materials, resulting in better combined powder particle size distributions and improved workability.

The regression models further show that higher apparent clay contents, as reflected with higher methylene blue values, resulted in increased HRWRA demand but decreased mini V-funnel time. Therefore, once sufficient HRWRA is provided to offset the effects of the clays, workability can be acceptable. Increasing the packing density of the microfines had only a slight effect on HRWRA demand but decreased mini V-funnel time to a much greater extent.

The results for concrete workability-shown in Fig. 10 and 11-matched those for mortar workability. The mixtures with microfines used as powder volume resulted in lower HRWRA demand and plastic viscosity than the mixtures with microfines used as aggregate volume, reflecting the differences in w/p and paste volume. When used as part of the powder volume, the TR-01 and DL-01 microfines resulted in increased HRWRA demand but unchanged plastic viscosity, reflecting the high methylene blue value of these aggregates. In contrast, GR-01 microfines resulted in increased HRWRA demand and plastic viscosity, reflecting the poor particle size distribution and shape characteristics of this material. The limestone microfines, which exhibited low methylene blue values and high combined powder packing densities, performed better than the other microfines.

Concrete hardened properties

Microfines generally had little to no effect on hardened properties when compared at a constant w/cm. The use of microfines resulted in slightly higher compressive strengths at both 24 hours (Fig. 12) and 28 days (Fig. 13) for a given w/cm. When microfines were used as part of the powder volume, the resulting reduction in cementitious materials content and increase in w/cm caused a reduction in compressive strength relative to mixture Control 1. When these mixtures were compared with mixture Control 2 with the same w/cm, the compressive strengths were slightly higher. The 28-day compressive strengths of approximately 35 MPa (5000 psi) for the mixtures with microfines used as powder volume are much more common in the concrete industry than the compressive strengths of approximately 60 MPa (8500 psi) for the mixtures with microfines used as aggregate volume. These results demonstrate how microfines can be used as part of the power volume to achieve the necessary w/p for workability and w/cm for hardened properties without significantly over-designing for hardened properties. It has been shown elsewhere that fine, noncementitious fillers23- including finely ground limestone in particular24-can accelerate the rate of hydration and increase the rate of earlyage strength gain. The contribution to hydration of the microfines considered in this study, however, was minimal because of the relative coarseness of the microfines in relation to the cementitious materials.

The use of microfines resulted in essentially no change in modulus of elasticity at a constant w/cm, as indicated in Fig. 14. It is generally accepted that modulus of elasticity is a function of the individual volumes and elastic moduli of the paste and aggregates and the size and quality of the interfacial transition zone between the paste and aggregates.1 In the data presented in Fig. 14, the w/cm and cementitious materials blend (25% fly ash and 75% portland cement by mass) appear to be dominant in affecting the development of modulus of elasticity for the use of microfines as part of the aggregate volume or powder volume. The w/cm and cementitious materials blend affect both the interfacial transition zone and the stiffness of the paste. In contrast, the microfines did not significantly contribute to the development of modulus of elasticity.

Microfines had little effect on flexural strength when evaluated at a constant w/cm (Fig. 15). It is likely that the microfines-due to their coarse size relative to all powder materials-had little effect on the interfacial transition zone and the paste-aggregate bond, resulting in little change in flexural strength for a given w/cm.

The use of microfines had no effect on rapid chloride permeability when microfines were used as part of the aggregate volume (Fig. 16). In this case, the increase in paste volume was likely offset by the reduction in w/p. When microfines were used as part of the powder, the rapid chloride permeability decreased by an average of 14% compared with the control mixture with the same w/cm. This reduction was likely due to the reduced total water content and w/p. In all cases, the rapid chloride permeability was low or very low based on ASTM C1202 qualitative ratings-due to the low w/cm, use of fly ash, and 91 days of moist curing prior to testing.

The use of microfines resulted in a slight increase in drying shrinkage for a given w/cm (Fig. 17). When microfines were used as part of the aggregate volume, the increase in paste volume could have contributed to the slight increase in drying shrinkage. When microfines were used as powder volume, the paste volume was unchanged and the drying shrinkage was generally lower than when microfines were used as part of the aggregate volume. Autogenous shrinkage was not independently measured.

The use of microfines as part of either the aggregate or powder volume generally resulted in reduced abrasion loss at a constant w/cm (Fig. 18). Any variation in abrasion resistance was likely related to the volume and abrasion resistance of the paste, as the fine and coarse aggregate sources were held constant. The w/cm and the blend of cementitious materials likely contributed to abrasion resistance to the extent they affected the development of paste microstructure. In addition, the microfines played a further role in enhancing abrasion loss, which was likely related to their hardness and their bonding within the paste. The improvement in abrasion resistance when microfines are used has been documented in other research.4,17

CONCLUSIONS

Based on the research presented in this paper, the following conclusions are reached.

1. Microfines can be used successfully in SCC. Microfines should be considered a powder material and accounted for as part of the paste volume. The w/p should be used for comparing workability and the w/cm for comparing longterm hardened properties;

2. When compared at constant w/p and paste volume, microfines typically increased HRWRA demand slightly and increased or decreased plastic viscosity;

3. When compared at a constant w/cm, microfines typically increased compressive strength at 24 hours and 28 days slightly, had no effect on modulus of elasticity and flexural strength, resulted in no change or a slight reduction in rapid chloride permeability, increased drying shrinkage slightly, and increased abrasion resistance; and

4. Increasing the specific surface area, span, and packing density of the microfines resulted in improved workability. The apparent presence of clay in microfines, as indicated with the methylene blue value, increased HRWRA demand but did not otherwise adversely affect the concrete.

ACKNOWLEDGMENTS

The research described in this report was conducted at the International Center for Aggregates Research (ICAR) at the University of Texas at Austin and was funded by the Aggregates Foundation for Technology, Research, and Education (AFTRE). The authors acknowledge C. F. Ferraris of the National Institute for Standards and Technology (NIST) for her advice and for providing access to laser diffraction particle size distribution measurements.