Cement Energy and Environment

) 1' signal decays. Higher values of a lead to smoother and more stable solutions, although broad with lower resolution. On the other hand, low values of a provide high resolution but with instability. Due to the high temporal resolution of measurements, smooth transition between time points is expected and random shifts of T2 peak positions indicate a suboptimal value of a. The value of a was chosen based on the L-Curve analysis [26] such that the temporal evolution of peak positions and shapes were smooth. Hence, narrow distributions, such as those presented in some studies (e.g., Muller et al. [21]), were not achievable under the requirement of smooth transitions between time points. 3.2. Overall trends and patterns Upon the commencement of measurements, following the addition of water (plus retardant) and mixing of paste, the water within the interstices between the clinker grains produced a well-defined T2 peak (T2L) positioned at - 20 ms that slowly shifted to lower values of T2 during the induction period before shifting rapidly to T2 values approximately an order magnitude lower (Fig. 4). This period of rapid decrease in T2 is termed here the 'Initial T2 Decline' (lTD). The appearance of a second peak at lower T2 values (T2S) following the lTD signified the commencement of gel pore formation . As hydration proceeded, the signal associated with the T2L peak decreased while that associated with the T2S peak increased (Fig. 5), indicating that capillary pore space was decreasing while gel pore space was increasing. This trend was broadly interpreted as indicating that capillary pore space was being replaced by hydrates containing gel pore space. Gel pore formation ceased approximately 30 to 40 h after commencing and, at the same time, capillary pore water content stabilized (Fig. 5). 1 H proton populations in hydrated OPC Ettrlngite l Solids Portland it~ Capillary pore waJtr Mobile tFr~ -r~J water Get pore water (C(l(l$1fill/led "'illt!r) lnterlayer water Mobile water measured by 1 H NMR relaxometry Incorporated water = Solids + lnterlayer water (StructunJ wal<lf) Total water • Mobile water + Incorporated water Fig. 3. Simple scheme for OPC, following the early hydration, illustrating the discrete populations of hydrogen proLOns resulting fmm the transfom1alion of added water a nd the formation ofhydration products. The C-5- H gel structure is based on the m lloid model ( 15,16J. Only the mobile water, comprising capH lary and gel pore water, was able to be measured by the NMR instrumenLTherefore, by measuring the tolal water at the start of the experiment, when inmrporated water (IW) equals zero, IW at aoy time was measured as the decrease in mobile water. The hydrogen proton populations discriminated in QENS analysis (9(, structural, constrained, and free water (shown in parentheses). correlate broadly with incorporated water. gel pore water, and capillary water respectively. , ... 20304ll50G0700 Time (h) Fig. 4. Positions of fitted peaks of the T 2 distributions (shown in Fig.2) during hydration of OPC with A) no retardant, B) 13 kg- 1 citrate, C) 2g ·1 citrate, or D) 10 mlkg Vertical dashed lines show the time t2, t2 and t3 that show the beginning of the ILD, gelpore formation, and C·S·H densification periods respectively. Time Fig. s . Data for the Relilrder N treaunent resulling from the llT and peak fitting procedures. before and after the applic<tlion of a Gaussian filter. illustrdte I ) the effect of th is smoothing operation and 2) the proposed phases ofearly hydration I hat are dearly d elineated by distinct evems. 1 1 - lnduclion period ends with the start of water consumption and the start of I he initial T 2 dedine (ITO). 1,- The ITO ends and gel pore fonnallon corrunences. 13 - Cel pore formation ends and C- S- H densification commences. 57