As high-performance concrete (HPC) has moved from the laboratory to field use, one problem sometimes encountered is its propensity for undergoing extensive self-desiccation and autogenous shrinkage . Due to the chemical shrinkage that occurs as the cement hydrates, empty pores are created within the cement paste, leading to a reduction in its internal relative humidity and a measurable shrinkage that may cause early-age cracking. This situation is intensified in HPC, relative to conventional concrete, due to its generally higher cement content, reduced w/c ratios, and incorporated pozzolanic mineral admixtures such as silica fume. The empty pores created during self-desiccation not only induce shrinkage stresses but also influence the kinetics of the hydration process, limiting the final degree of hydration, and thus strength, that can be achieved relative to that obtainable under saturated curing conditions .
This "curing" problem was recognized nearly ten years ago by Philleo , who suggested incorporating saturated lightweight fine aggregate (LWFA) into the concrete mixture to provide an internal source of water needed to replace that consumed by chemical shrinkage during hydration (curing). As the cement hydrates, this extra water will be drawn from the relatively "large" pores in the LWFA into the much smaller ones in the cement paste. This will minimize the development of autogenous shrinkage as the shrinkage stress is controlled by the size of the empty pores, via the Kelvin-Laplace equation . Unfortunately, little was done to follow up on Philleo's idea until the mid-1990's when Weber and Reinhardt  once again proposed the use of saturated lightweight aggregates to support the curing of concrete. In their experimental program, replacing a portion of the fine aggregates by their saturated LWA counterparts resulted in concretes which were "considerably less sensitive to the curing process" . More recently, Bentur and his colleagues [4,5] have prepared high-strength concretes with a mixture of LWFAs and normal weight aggregates and observed that the initial autogenous shrinkage measured for HPC can be eliminated by the judicious replacement of a portion of the fine aggregates with either saturated or air-dried LWFA. In this communication, equations are derived to estimate the replacement level necessary to avoid autogenous shrinkage as a function of mixture proportions. Additionally, a 3-D concrete microstructural model is applied to determining, for various replacement levels and aggregate gradations, the fraction of the hydrating cement paste within a given distance of the LWFA surfaces. We focus on replacement of the fine aggregate component as opposed to the coarse, due to both strength considerations and the fact that the much higher surface area of the fine relative to the coarse aggregate will result in a more uniform distribution of this additional curing water within the 3-D concrete microstructure. The results of the 3-D microstructural model are compared with an approximation based upon the equations developed previously by Lu and Torquato  for the generic hard core/soft shell microstructural model. This paper only considers the availability and spatial proximity of the additional water introduced via the LWFA and does not address the processing (rheology) issues which may need to be addressed to offset the effects of the generally rougher surfaces of the LWFA on the flow properties of the concrete.