Mixture Workability

Workability, a property which is often overlooked in literature, is defined to be “the ease and homogeneity with which concrete can be mixed, placed, consolidated and finished” (ACI 116.R, 1990) in its fresh state. Workability is a very important measure of performance, since the fresh state properties of concrete ultimately affects its hardened state properties as well (Mindess et al., 2002). For example, mixtures with poor workability may leave voids in the formwork (known as honeycomb), be improperly compacted, exhibit segregation or excessive bleeding, all of which can affect the long term strength and durability of concrete (Mindess et al., 2002). Typically, the level of workability required for a concrete mixture depends on the application. Since the research presented here does not focus on one particular application, the level of workability desired for the pozzolan pastes in this research were the same as that of the control mixture without any pozzolans.

The water requirement test of ASTM C 618 (2012) provides a simplistic assessment of a pozzolan’s effect on mixture workability. However, it cannot measure the actual rheological properties of the mixture, like yield stress and viscosity. The viscosity of a material can be defined as a measurement of the material’s internal resistance to flow. More specifically, viscosity refers to the amount of shear stress that is needed to increase the shear rate or the velocity gradient of the fluid flow cross-section by one unit (Mindess et al., 2002). On the other hand, the yield stress of the material is the shear stress that must be exceeded before the material can start to flow (Mechtcherine et al., 2014). Having a thorough scientific understanding of these properties is crucial when trying to optimize fresh state properties of concrete mixtures or finding ways to mitigate the high water demand of materials like zeolites. As such, in this research, mixture workability was evaluated using an MCR 301 Anton Paar rotational rheometer than can

measure the rheological properties of cementitious pastes. A cup and bob measuring system geometry was used, with a 1.0 mm gap between the bottom of the cup and the bob. Similar to the mixture design for isothermal calorimetry and TGA, the control paste (with no SCMs) was made with 50 g of cement and 22.5 g of water, giving the paste a w/cm of 0.45. For the pozzolan pastes, 20% of the cement by mass was replaced with the pozzolan being tested. The water content was kept the same as the control mixture, giving the pozzolan pastes a w/cm of 0.45 as well. The pastes were mixed mechanically for 2 minutes using a Caframo Compact Digital BDC 2002 overhead stirrer at 1000 rpm. Approximately 19 mL of the mixed sample was added to the cup for rheological testing. Prior to each test, the pastes were pre-sheared for 4 minutes at a shear rate of 50 s-1. This was done to reduce the effects of shear history on the samples and to ensure a similar starting point across all tests. After pre-shearing, the samples were allowed to rest for 30 seconds. Then, the shear rate was gradually increased from 10 s-1 to 50 s-1 and then brought back down to 10 s-1. Both the increase and decrease in shear rate were done in increments of 10 s-1. Furthermore, the shear rate was held constant for 3 minutes after each increment to ensure that an equilibrium state had been reached. Ten data points were used in each equilibrium range to determine the average shear stress of the paste at each shear rate. The shear stress and the shear rate obtained from the test are then used to graph a rheological flow curve, from which the viscosity and yield stress of the mixture can be determined by fitting different models. The total time for the rheology test (including the time to pre-shear) was 31.5 minutes. For each type of paste, the entire test was repeated at least two times.

In the current study, cementitious mixtures that showed a linear trend in their flow curves were analyzed using the Bingham model, a popular method to describe the rheological behavior of cementitious mixtures (Mechtcherine et al., 2014; Roussel et al.,

2007). Using a linear trend line to fit the flow curve data, the Bingham model defines the slope of the linear trend line to be the viscosity, and the y-axis intercept to be the yield stress (Mindess et al., 2002). For mixtures that showed a non-linear trend in their flow curves, instead of using the Bingham model, the average apparent viscosity of the mixture was calculated from ten data points measured in equilibrium range of each shear rate. Apparent viscosity is defined to be “the quotient of shear stress divided by rate of shear when this quotient is dependent on the rate of shear” (Tattersall and Banfill, 1983). In other words, unlike the Bingham viscosity, which is independent of the shear rate applied to the mixture, apparent viscosity refers to the viscosity (or the ratio of shear stress to shear rate) of a mixture at a particular shear rate. Finally, it must be noted that all the rheological analysis was based on the region of the flow curve where the shear rate was being gradually decreased from 50 s-1 to 10 s-1, as the data is considered to be more stable when the shear rate is being decreased, rather than when it is increased.