Abstract
"Understanding the transport properties of water in cementitious materials is crucial in predicting the material’s service life. The water and gas diffusivity depends on the cement phase composition, pore structure, microcracks and degree of water saturation. However, it is not fully understood why water ingress into cement paste deviates from the linear dependency with the square root of time. That relationship is true for most other porous materials, where sorptivity progresses according to the Fickian law. Several mechanisms such as shrinkage and swelling, creep, drying-induced micro-cracking, delayed hydration, calcium hydroxide precipitation, or a time-dependent permeability constant have been proposed to explain the anomalous transport in cementitious materials.
In the present study, 1H NMR (nuclear magnetic resonance) relaxometry and imaging techniques have been used to investigate the evolution of porosity distribution during capillary sorption experiments. Cement paste samples were dried to empty the gel pores partially or fully. In all samples, dynamic cement paste microstructure was observed during re-wetting experiments. That behaviour may partially be responsible for anomalous water transport in hardened cement paste.
A modified transport model introduced by McDonald and co-workers in a publication [1] was used to fit spatially resolved SPRITE (single-point ramped imaging with T1-enhancement) MRI (magnetic resonance imaging) data of cylindrical 60 mm long samples exposed to different drying times. The drying duration and severity affect the porosity relaxation time constants and relate to how firmly C-S-H sheets collapsed during the pre-drying process. Results revealed that the duration of porosity relaxation to the pre-drying state occurs on different timescales ranging from days to months. It is an important discovery, especially given that in smaller 1.5 mm samples, the timescale is an order of magnitude lower. Contrary to the micro-scale, macro-scale experiments investigate the influence of factors such as drying-induced microcracks and the overall pore network tortuosity. These factors seem to be the reason for dramatically increased water diffusivity in 60 mm long samples.
The second aim of this study was to find a link between liquid and gas diffusivity in non-carbonated and carbonated hardened white cement pastes with a water-to-cement ratio of 0.5. The specially designed cuboid shape of the samples enabled homogeneous carbonation and capillary rise experiments. Additionally, the influence of ground granulated blast-furnace slag (ggbs) on transport properties was investigated. CPMG (Carr-Purcell-Meiboom-Gill) studies have shown reduced overall water-filled porosity in carbonated samples with and without 50 % slag addition. However, the water-filled porosity decrease was more pronounced in blended pastes. Nonetheless, gas and liquid diffusivity in white cement paste decreased upon carbonation. The opposite effect was observed in samples with 50 % slag addition.
Additionally, a validation of the transport model based on the Monte-Carlo (MC) diffusion simulation is presented. Various assumptions and parameters were tested to investigate the pore relaxation mechanisms. The best agreement with the experimental data was found for the model with the conserved gel porosity, leading to only one porosity relaxation time constant.
Further analysis is presented in an epilogue involving a model with one relaxation time constant."