Cement Energy and Environment

distinct events in water transformation with the progress of hydration reactions. The low field (2 MHz) NMR technique used in this study detected signals from only the mobile water contained in the capillary and gel pores with water incorporation into hydration produ:::ts highly correlated with heat production, including the reproduction of subtle features. Following the induction period , an initial T2 decline pe!riod, which encompassed the acceleration period of hydration, was characterized by incorporation of water into hydration products without any associated gel pore formation. Gel pore formation commenced at the peak in hydration rate, indicating a change in the morphology of C-S-H growth. All the observed features of microstructural development were preserved under retardation. 1.0 Introduction Development of microstructure during the early hydration of Ordinary Portland Cement (OPC) has attracted interest both due to the importance of microstructure to emmgent bulk properties and to the advancements in analytical tools and techniques suitable for probing this complex evolving system [1] and [2] . Tt1e apparent potential for influencing microstructure development by controlling hydration reactions has created particular interest in links betw1aen the two processes, however, these links, and the scale of microstructure (including molecular structure) at which investigations should optimally focus, remain elusive and are lacking in mechanism– based analysis [3]. Non-invasive methods of continuously monitoring reactions and changes in hydrating cement are highly attractive for obvious reasons. Detection of the state of hydrogen protons, and discrimination and quantification of their various populations, has proved to be a highly fruitful approach since 1) all hydrogen protons in the system originate from added water, 2) water is a key player in hydration reactions, and 3) the state of any hydrogen proton is a function of its chemical and physical environment including microstructure. A number of techniques have been used successfully to probe the state of water in hydrating cement including broadband dielectric spectroscopy (BDS)[7] and [8], quasi-elastic neutron scattering (QENS) [9] and [1 0], and nuclear magnetic resonance (NMF~) [11], [12] and [1 3]. 1 H NMR relaxometry has proved to be a powerful technique that is relatively easy to implement, especially with the development of bench-top instruments capable of rapid analysis of large samples. In combination with the development of a robust conceptual model for the microstructure of C-S-H gel (14], [15] and (16], significant advances have been achieved over the last decade in the design of 1 H NMR relaxometry experiments, the interpretation of data, and the identification of discrete proton populations associated with proposed microstructural features [17], [18], [19] and [20]. Using quadrature (solid) echo and Carr– Purceii-Meiboom-Gill CPMG (spin) echo techniques in combination, Muller et al. [21] identified four key populations of hydrogen protons in hydrating white cement: 1) hydrogen protons associated with the minerals ettringite and portlandite, 2) C-S-H interlayer water, 3) gel pore water associated with the small interstices between C-S-H globules, and 4) capillary pore water initially associated with the space between the cement grains and , later, as inter-hydrate spaces. Of these four populations, identification of the C-5-H interlayer water is probably least resolved , likely due to the presence of both chemically and physically bound water (12]. Also, bench-top NMR instruments are often limited to a minimum echo time in the vicinity of 100 IJS, which is the same as that reported for the position of the T2 (transverse relaxation time) peak associated with C-S-H interlayer water [11] and [21]. Information regarding the populations of low-mobility hydrogen protons with extremely fast relaxation times has only been achieved using white cements, free from paramagnetic impurities. The impact of routinely used and cost effective admixtures on cement hydration processes linked to microstructural development has received limited attention. High levels of sucrose have been shown to have an impact on pore structure and spatial distribution of hydrates, suggested to be due to enhanced homogeneous nucleation of C– S-H in the pore space away from the cement grains [4]. More recently, synchrotron-based XRD enhanced microtomography has been used to reveal a profound effect of superplasticizers on the distribution of C-S-H in hydrated OPC. In the absence of superplasticizer, C-5-H distribution was highly correlated with that of the cement grains, while in the presence of superplasticizer, 54