The efficiency of one fluid displacing another in a permeable medium depends on the pore-scale dynamics at the main wetting front. Experiments have shown that the frontal dynamics can result in two different flow regimes: a sharp and a diffuse front. In the sharp front regime, the displacing fluid occupies nearly all the pores and throats behind the main wetting front and the saturation changes abruptly. In contrast, in the diffuse front regime, pores are filled gradually at the main wetting front, and the saturation change is gradual in space. The different fronts can greatly alter the relative permeability curves, the trapping mechanisms, and the displacement efficiency. Directly measuring the sharpness of the front is difficult. Instead, here we correlate the front sharpness to saturation overshoot, which occurs for moderate to high flux vertical displacements of low density fluid by a higher density fluid in 1-D homogeneous permeable media. We hypothesize the sharpness of wetting front can be explained by competition between two different pore - filling mechanisms (called snap-off and piston-like) with the competition controlled by the velocity of the front and thus the injected flux. We conduct series of infiltration experiments to determine the saturation profile as a function of flux for seven different fluids. We find that for each fluid there is a flux (called overshoot flux) below which saturation overshoot ceases and the front is diffuse. We find that the overshoot flux depends inversely on the invading fluid’s viscosity, and shows little or no dependence on the invading fluid’s surface tension, vapor pressure, and its miscibility with water / text

The efficiency of one fluid displacing another in permeable media depends greatly on the pore-scale dynamics at the main wetting front. Experiments have shown that the frontal dynamics can result in two different flow regimes: a stable and an unstable front. In stable displacements, any perturbation of the front will diminish with time and the effect of variation in permeability will be lessened. In contrast, in unstable displacements any perturbation of the front will grow with time and any variation in permeability will be magnified. In this dissertation, the stability of two different displacement processes are contemplated; a) vertical infiltration of dense liquid into dry sand from above and b) horizontal displacement of nanoparticle suspension with high pressure liquid CO₂. Significant insights are obtained by measuring the in-situ flow patterns in real time with a light transmission method and CT scanning. Vertical infiltration of dense fluid into dry sands from above is often observed to be unstable and produce gravity driven fingers. The formation of gravity fingers can have large consequences on the sweep efficiency of a displacement. Infiltration experiments showed that gravity driven fingers have a unique saturation profile known as saturation overshoot with a higher saturation at the finger tips than the saturation at the finger tail. Despite the vast number of theoretical and experimental investigations, conditions under which the front is unstable, remain unclear. To determine what controls the saturation overshoot and how it relates to the dynamics at the initial wetting front, saturation overshoot was measured as a function of flux for seven different liquids. These liquids gave a range of molecular weights, viscosities, and vapor pressures. It is found that for each fluid there is a flux (called overshoot flux) below which saturation overshoot ceases and the front is diffuse. The magnitude of the overshoot flux depends inversely on the invading fluid's viscosity and shows little or no dependence on the invading fluid's surface tension, vapor pressure, or miscibility with water. Since the saturation overshoot is not described by the continuum multi-phase flow models, the experimental results are used to develop a semi-continuum model that bridges the continuum-scale and pore-scale physics. The proposed model predicts the observed dependence of overshoot on media permeability and invading fluid properties. At the planned depth for CO₂ injection, either as an enhanced oil recovery technique or for CO₂ storage, CO₂ is typically less dense and less viscous than the in-situ fluid. Therefore, CO₂ injection is unstable and produces viscous fingers. This can greatly reduce the efficiency of a CO₂ flood or CO₂ storage capacity of an aquifer. To remedy this behavior, surface treated nanoparticles were used to reduce the mobility of injected CO₂. Displacement experiments were performed at low pressure with a CO₂ analogue (n-octane) fluid and at high pressure with liquid CO₂. Saturation distributions and pressure drops were measured in real time with the CT scanner when high pressure liquid CO₂ or n-octane was used to displace brine in different cores with and without suspended nanoparticles. In the presence of nanoparticles, the displacement front is more spatially uniform with a later breakthrough compared to the same experiment with no suspended nanoparticles. These observations suggest that nanoparticle stabilized foam, which forms during the displacement, acts to suppress the instability. It is argued that the generation of droplets occurs at the leading front of all drainage displacements. In the presence of nanoparticles, these droplets are preserved when nanoparticle adhere at the fluid-fluid interface. The new mechanism for foam generation described here, provides an interesting alternative for mobility control in CO₂ floods. Moreover, the same mechanism can potentially a) increase the CO₂ storage capacity of an aquifer, b) enhance the CO₂ capillary trapping, and c) provide an engineered barrier to CO₂ leakage from a storage sites, thereby alleviating the risk of contaminating the overlying fresh groundwater resources for CO₂ storage projects. / text