Distribution of Pore Sizes in White Concrete

  • PDF / 392,578 Bytes
  • 7 Pages / 414.72 x 648 pts Page_size
  • 90 Downloads / 199 Views

DOWNLOAD

REPORT


ABSTRACT NMR is used to study the evolution of the discrete distribution of proton magnetization fractions in hydrating synthetic white cement. The results at 100 hours hydration time are modelled using a trimodel pore size distribution. A good correspondence is found between the NMR analysis and SANS results reported in the literature.

INTRODUCTION During cement paste hydration some water reacts chemically with cement clinker while the remaining water, being liquid-like, resides in pores and voids of the hardening cement paste. The hydrogen atoms on liquid-like water and hydrogen atoms (initially also from water molecules) in the cement's solid structures are in several chemically, structurally, and dynamically distinct environments. In this research we use proton nuclear magnetic resonance (NMR) to study the establishment and evolution of these environments in hardening cement paste. NMR is ideally suited for this purpose because it uses hydrogen nuclei as probes and monitors many details about the evolving structure and dynamics of these environments at the molecular level. All proton spins in a particular environment (repeated throughout the cement paste) are considered to constitute a spin "group" which is characterized by its spin-lattice relaxation time (T,), its spin-spin relaxation time (T2), and its magnetization fraction (MO,). The T1 and T2 yield dynamical information about the proton bearing molecules in the spin group. While T2 generally decreases as the molecular environment becomes more rigid, T1 is normally at a minimum when the molecular dynamical frequencies are equal to the NMR Larmor frequency employed (in our case 40 MHz). The proton magnetization fraction is exactly proportional to the number of protons constituting the spin group, thus it measures the proton group's relative size. In NMR studies of heterogeneous materials, such as hydrating cement paste, two main analysis protocols are used. The first involves the characterization of the system with a discrete distribution of spin groups; e.g. a sum of delta functions. Different spin groups represent distinct structural components of the cement paste (e.g., C-S-H, CH, other crystallites such as ettringite, and water in several kinds of pores). The resolution of the CH magnetization is remarkable. The C-S-H and other crystallites are resolved relying in part on stoichiometry [1]. The second method involves the fitting of a continuous distribution of different environments (such as water in layers, pores, capillaries, etc.) within the structure of the cement paste. In this case the spin-lattice recovery function is characterized by a continuous distribution of T, values. Indications are that both approaches are applicable in hydrating cement paste as the magnetization recovery function in this material has been found to be strongly nonexponential. On the one hand, the nonexponentiality is such that its resolution into several identifiable components by spin-grouping is well justified [1] at hydration times beyond - 50 h if the recovery of the