Cement Energy and Environment

the C-S-H distribution was random with homogeneous nucleation of C-S-H apparently occurring in the pore spaces [5]. A greater extent of homogeneous nucleation might be expected to lead to a more uniform microstructure with less capillary porosity and a higher specific surface area [6], with implications for mechanical bulk properties and permeability. In this study we have monitored, with high temporal resolution, early hydration of OPC with and without retardant using bench-top 1H NMR relaxometry and isothermal calorimetry with the aim of 1) determining which populations of water are measurable by bench-top NMR in hydrating OPC and how their dynamics relate to heat output from hydration reactions, and 2) investigating whether retardation of early hydration reactions impacts on the dynamics of water populations, and associated microstructure, in a measurable manner that suggests an evolution towards an altered microstructure, such as more homogeneous C-S-H distribution, with implications for resultant bulk properties. 2.0 Materials and methods 2.1 Materials Ordinary Portland cement (OPC) was obtained from Boral Limited, Australia. The elemental analysis is shown in Table 1. Citrate was prepared as a 10% solution of sodium citrate (trisodium citrate dihydrate, Chern Lab). A commercial retardant commonly used in concrete manufacture (Retarder N, Sika; recommended typical dosage of 2 ± 1 mLkg- 1) containing selected carbohydrates in liquid form was also used in investigations. All solutions were made using MilliQ (MQ, Millipore) water (> 18 MO em). MQ water was used in the production of cement paste. 2.2 Methods All investigations used cement pastes mixed at a water-to-cement ratio of 0.45. Retardant solutions were added to mixing water in the desired proportions. Retarder N was added on a volume per weight of cement basis (mL kg- 1), while citrate was added on a weight per weight of cement basis (g kg-1). Pastes were produced by adding cement to mixing water in a plastic beaker and stirring vigorously with a spatula by hand for two minutes. Excellent reproducibility of the mixing, atleast with regards to the progression of hydration, was confirmed from calorimetry analyses (data not shown). For NMR experiments, paste was mixed and immediately poured into a 45 mL styrene vial sealed with a screw top lid, which was then inserted into the instrument bore to commence measurements. For calorimetry, 7 g of mixed cement paste was weighed into a glass vial (2 replicates), crimp sealed and inserted into the instrument. Table 1 Main elemental compositioo of investigated Ole OXide Weight% CaD 64.56 Si02 20.09 Al203 5.38 Fe203 2.96 503 2.67 MgO 1.22 Transverse relaxation times (T2) for hydrating cement were measured using a Carr-Purceii– Meiboom-Gill (CPMG) pulse sequence [22,23] in a Magritek Rock Core Analyser operating at 2 MHz. For each CPMG dataset, 2000 echoes were acquired with an echo time of 105 iJS, resulting in a total relaxation delay of 210 ms. Temperature within the magnet bore was maintained at 23 oc throughout the period of measurement. Measurements commenced immediately after sample insertion and were undertaken every 20 min for up to 80 h without disturbing the sample. T2 distributions were generated via an inverse Laplace transform (ILT) using an exponential kernel [24,25] where the regularization parameter a was optimized according to the L-Curve analysis [26] of the data for intermediate hydration times. Log-normal fits were used to deconvolute the distributions and discriminate the contributions from two hydrogen proton populations : 1) T2L– Iong relaxation times, and 2) T2S-short relaxation times (Fig. 1). The total amount of liquid water in the system was quantified at the start of the experiment from the amplitude of the ILT multi– exponential fitting curve at the time of the first echo. Consumption of 'measurable' water as hydration progressed was monitored using this metric. The weight of the sealed vial containing added cement and water was measured at the start and end of the experiment in order to confirm that no water was lost from the system due to evaporation. The contributions of both populations 55