Restoration of Groundwater Quality Following Pilot-Scale Acidic In-Situ Uranium Leaching at Nine-Mile Lake Site Near Casper, Wyoming

Sandstone formations contain a variety of minerals including carbonates, clays, feldspars, quartz, and, in some cases, valuable compounds of uranium, vanadium, and copper. When a reactive fluid such as acid is injected into these porous rocks to dissolve the minerals, their dissolution rates are different because of differences instructure and chemical composition. Consequently multiple reaction zones develop as a reactive fluid flows through the rock. This study, describing the movement of these reaction zones during sandstone matrix acidizing, also has application in hydrometallurgy and in-situacid leaching of minerals. Laboratory tests to predict the effects of acid injection usually are carried out by flowing acid through a coresample while permeability and fluid effluent concentration are monitored. If the minerals in a particular sample dissolve at approximately the same rate, a lumped parameter model satisfactorily describes the movement of a single reaction zone through the core. However, for the case of high temperatures, long cores, or long injection times, this model does not predict the effluent acid concentration correctly, and a distributed parameter model including the movement of multiple reaction zones is required. This paper describes the development of such a model and its verification by core flowexperiments. The study shows that when various minerals including quartz are being dissolved during matrix acidizing, larger acid volumes are required than predicted byprevious models. Introduction The dissolution of clays, feldspars, and other minerals ina sandstone can be achieved by the injection of mixtures of hydrochloric and hydrofluoric acid. Because of differences in chemical composition and structure, the acid dissolution rates of these minerals are different. As aresult, multiple reaction zones are developed as the acid flows through the porous sandstone. An earlier investigation demonstrated that if we assume that quartz dissolves so slowly in hydrochloric/hydrofluoric acid (mud acid) as to be considered unreactive and if the other minerals dissolve atapproximately the same rate, a lumped parameter model satisfactorily describes the movement of a single reaction front or zone. In this model, reactive minerals such as various clays and feldspars are lumped together and treated as a single dissolvable species. Carbonates are assumed dissolved in hydrochloric acid and are not involved in the hydrofluoric acid reaction. For high temperatures, long injection times, and long sandstone penetration distances (as in long test cores), this lumped parameter model does not predict correctly the changing acid concentration and the movement of multiple reactionzones during acid injection. A distributed parametermodel is necessary to describe the acid stimulation of sandstones when the difference in acid dissolution of the various minerals is taken into account. This paperdescribes the development of such a model. In addition, it shows how this model, coupled with the analysis of laboratory core flood experiments, predicts the radial movement of reaction zones and corresponding changesin the permeability during acid stimulation. Acidizing Linear Sandstone Cores Evaluation of the DissolutionParameters From Effluent Data Sandstone stimulation studies in the laboratory generallyare carried out by flowing hydrochloric/hydrofluoricacid mixtures at constant rate through cylindrical coresamples and monitoring the permeability changes and the effluent acid concentration. The lumped parameter model was developed to describe the movement of the acid reaction front through the core. SPEJ P. 99^