Rheology of cement mortar and concrete as influenced by the geometric features of manufactured sand

Manufactured sand (MS) has attracted increasing attention and been gradually used to replace natural sand as aggregate of mortar and concrete, due to the decreasing availability of natural sand and environmental impacts caused by over exploitation. The crushing process provides MS with irregular particle shape and rough surface texture. These features make the properties of mortar and concrete with MS different from those with natural sand, especially in terms of rheological properties. Evaluation of the rheology of mortar and concrete with MS is not only a key step to ensuring proper flow and successful casting in formworks, but also provides guidance for mix design of concrete mixtures and production of MS. This thesis addresses a study about the effects of geometric features of MS on the rheology of mortar and concrete mixtures. The geometric features of MS particles were analyzed in terms of overall shape, angularity and roughness from random sections of particles. Various MS particles with different geometric features and particle sizes were used to prepare cement mortar mixtures, of which the rheological properties were measured by means of a rheometer. The relationship between the rheology of mortar mixture and the characteristics of MS was analyzed from the viewpoints of packing properties of MS, paste film thickness (PFT) and relative paste film thickness (R_PFT) of mixtures. A new method for measuring the specific surface area (SSA) was proposed for the purpose of calculation of PFT and R_PFT. This relationship at mortar level was then extended to concrete level to predict the rheology of concrete based on a multi-level biphasic approach. A 2D image-based methodology was proposed to measure the SSA of irregular aggregate by random sectioning. Factors that may influence SSA result were analyzed. Conventional methods including spherical assumption, BET and computed tomography (CT) tests were used and compared in this study. Results show that spherical assumption provides the lowest SSA among these methods since the feature of anisotropy in dimensions is not considered. The BET method yields SSA values one order of magnitude higher than others, which is attributed to the fact that the BET method measures each position of particles that nitrogen molecule can be adsorbed on during the applied relative pressure, based on the ‘pixel’ of nitrogen molecule. The proposed random sectioning method presents very similar SSA result compared to CT method, indicating that it can be considered as a reliable method. To improve the reliability of SSA by random sectioning method, the number of samples should be high enough to reach a constant result and the thresholding algorithm should be adequate. Besides, a higher pixel resolution provides a higher SSA value. The comparison among these methods demonstrates that it is necessary to determine the scale at which features of the surface are supposed to be captured before selecting the optimal testing method. The geometric features including overall shape, angularity and surface roughness of MS particles were analyzed by aspect ratio, convexity area ratio and convexity perimeter ratio respectively. Circularity was also employed to comprehensively evaluate the geometric features of MS particles by measuring the similarity between the particle and the perfect sphere. The number of tested particles for steady geometric parameters was analyzed and it is found 600 particles are sufficient for the measurement of geometric features of particles used in this study. All MS particles in this study show similar angularity results. The roughness of each MS increases as the particle size increases while the circularity decreases with the increase of particle size. For aspect ratio, different sands show diverse trends with particle size. The plastic viscosity and yield stress of mortar mixtures was calculated based on the Bingham model. In terms of relative plastic viscosity, the Robinson model shows optimal fittings for all mixtures and is thus used to determine the packing fraction of MS under shearing. From the particle packing viewpoint, shear-induced orientation increases the packing fraction of non-spherical MS particles from the random loose packing fraction. This influence is increasingly prominent with the decrease of circularity. The relative volume fraction is an important parameter influencing the relative plastic viscosity of mortar mixtures with MS whereas the relative paste film thickness (R_PFT) is found as the dominating factor. The dependence of plastic viscosity of cement mortar on geometric features and particle size of MS can be attributed to their influences on the packing fraction and surface area of MS particles. Similarly, the relative yield stress of mixtures is dominated by R_PFT as well. While comparing the fitted parameters from the relative plastic viscosity and relative yield stress, it is found that the mixture with spherical glass beads (GB) has a higher relative plastic viscosity than its relative yield stress while the mixtures with non-spherical MS particles have slightly higher relative yield stress than relative plastic viscosity. Moreover, the perceived effect of MS on the plastic viscosity of mortar mixtures is influenced by water to cement (W/C) ratio. Increased W/C ratio slightly intensifies the influence of R_PFT on relative plastic viscosity. Nevertheless, the dependence of relative yield stress of mortar mixtures on R_PFT is independent of W/C ratio. The relationship between the relative rheology of mortar mixtures and R_PFT is extended to concrete level. The relative rheology of concrete is considered dependent on the relative mortar film thickness (R_MFT). Then the rheology of concrete with MS is predicted based on a multi-level biphasic approach. Concrete is considered as a two-phase suspension of coarse aggregates in the suspending matrix of cement mortar. The relative rheology of concrete to mortar mixtures is predicted from R_MFT. At mortar level, the mixture is taken as a suspension with MS particles as inclusions and cement paste as the matrix. The relative rheology of mortar mixture to cement paste is predicted based on R_PFT. According to these findings, one may predict the rheology of mortar and concrete with MS. Besides, this approach can also be used for the mix proportioning of concrete, especially self-compacting concrete while rheology needs to be specifically concerned. In terms of optimization and manufacturing of aggregate, spherical granules with smooth surface are preferred from the rheological viewpoint