Cement Energy and Environment

designated as interlayer water in the Colloidai-Model-11 of C-S-H [15]. Hence, there is some uncertainty within the literature re9arding the chemical and physical status of this interlayer water and its role in heat producing hydration reactions. Quasi-elastic neutron scattering (QENS) studies of early cement hydration support the proposal that interlayer water participates in heat producing reactions. Refining an earlier approach [33], Thomas et al. [9) decomposed the mobile water signal into free and constrained water fractions, and combined constrained water with structural (solid) water fractions to form a bound water index (BWI). While the BWI was able to be fitted well with a hydration model [34], structural water alone correlated to heat production durin!~ cement hydration and formation of constrained water ceased after -1 day despite continuation of hydration [9]. These authors concluded that structural water in C-S-H included both chemically bound and interlayer water and that constrained water included water in the smallest of gel pores and adsorbed to solid-pore interfaces. Constrained water was contended to be solely associated with high surface area, low density (LD) C-S-H (LD C-S-H) , while both LD (low density) and HD (high density) C- S- H contain structural water. Despite an uncertain role of constrained water in early hydration, the conceptual model of interlayer water developed by these QEI\IS studies tends to support our proposal that IW includes interlayer water and that the formation of interlayer water is part of the heat producing hydration reactions. The inclusion of gel pore water in the BWI during the main hydration peak highlights the close connection between hydrate formation and gel pore formation. The correlations between the populations identified using QENS and those proposed in this study are illustrated in Fi!l 3. Other studies have included water with greater mobility than interlayer water in hydration reactions [35,36] and excluded interlayer water from reactions [37], using differential scanning calorimetry and the T1 signal amplitude from 1 H NMR relaxometry respectively. In studies where XRD was used to quantify the consumption and formation of different phases during hydration of alite [38] and cement [39], the relationships between enthalpy of reaction and water content of the hydration products implied the involvement of interlayer water in heat producing reactions. Using a simple set reactions based on a formulation for C-S-H with a H:Si ratio of 2.6, these XRD measurements enabled theoretical calculation of heat production which matched calorimetric measurements well. This H:Si ratio was well in excess of 1.8, the value recently considered to account for C-S-H associated water, including interlayer water [16]. While it appears that heat production may possibly be correlated with populations of water that both include and exclude interlayer water, on balance the weight of evidence supports the inclusion of interlayer water in heat producing reactions and supports the proposal that it is part of the IW fraction in this study. 3.5 Correlation of rate of incorporated water production with heat flow The rate of production of IW is very closely correlated in all cases with the rate of heat production as measured by calorimetry (Fig. 7). While in the absence of retardant the rate of IW production drops below that of heat flow during the first part of the deceleration phase (possibly related to C3A dissolution as discussed further below), the reproduction of the small peak in heat flow at -37 h is quite notable (Fig. 7A). In the citrate treatments, this feature is still evident in heat flow, as a broad peak around 43 h and 51 h respectively, and again appears to be reproduced in IW production (Fig. 78 and C). Interestingly, these subtle features appear to be accentuated by both the rate of disappearance of capillary pores and the rate of formation of gel pores, which tend to mirror each other. Even in the Retarder N treatment (Fig. 70), where this small secondary feature has not been detected by either heat flow or IW production measurements, there appears to be a similar pair of mirrored peaks. The relationship between capillary and gel pores, and the possible interpretation of these mirrored peaks will be discussed further in Section 3.7. The proportionality between the rate of heat production and the rate of IW production was identical for all treatments (note that the same axes were used for all plots), even where high retardation has slowed the reaction rate, further supporting the proposal that the IW fraction represents the water taking part in the heat producing reactions. Also, proportionality was maintained throughout early hydration, suggesting 60 ,.