Abstract
In this thesis the Monte Carlo translational diffusion simulation method is used to calculate 1H (water) nuclear magnetic resonance [NMR] spin relaxation rates in model cement pores. The simulation results are used to gain insight into the validity of analytical models of water NMR relaxation, and hence water dynamics, in cement pastes. A series of simulated systems of increasing complexity are studied beginning with pure two- and three-dimensional bulk water for validation of the numerical algorithms against known results, through a study of bulk water confined to a quasi two-dimensional planar pore and onto a study of water in quasi-two dimensional pores of varying size with restricted surface dynamics and/or surface paramagnetic interactions. This last system mimics the widely applied Korb model of relaxation in cements and allows a novel means to study underlying assumptions of averaging in the biphasic model of relaxation in porous systems. Throughout the work, simulations are parameterised using the results of molecular dynamics simulations obtained in a parallel study by J. S. Bhatt (University of Surrey). The Monte Carlo simulations are carried out over microsecond durations allowing for the determination of the NMR relaxation rates at experimentally used low frequencies of the order of 1-20 MHz. The NMR relaxation rates are calculated from the spin-pair dipolar autocorrelation functions, G*(t), evaluated from the molecular trajectory records from the Monte Carlo simulations. The G*(t) are described in terms of characteristic times associated with the dynamical processes performed by the H nuclei in the quasi-two dimensional pores, and are compared with new analytic forms calculated by D. A. Faux (University of Surrey). The displacement of the paramagnetic impurities from the surface adsorbed layer of water is demonstrated to be an important parameter. In conclusion, this work demonstrates the validity of Faux's analytical results that in turn demonstrate the limitations of Korb's model of relaxation in cements and in particular Korb's erroneous conclusion of a long surface residence time for water adsorbed to cement pore surfaces.