Summary.
A study of the reaction of alkaline chemicals with minerals constituting reservoir rock is presented. Static tests were conducted with high concentrations of NaOH and orthosilicate solutions and minerals (montmorillonite, kaolinite, illite, and quartz sand). The reaction time varied from 10 minutes to 2 months. Solutions were analyzed for hydroxide, Si, and At, and solids were analyzed by X-ray diffraction (XRD) and other spectroscopic techniques. The loss of useful alkalinity was the highest with kaolinite and the least with quartz sand for high alkali concentration (5 wt%) or temperature (180F [82C]). A detailed study of kaolinite/alkali reaction kinetics and quartz/alkali equilibrium is presented. presented.
Introduction
Extensive and fast propagation of hydroxide ions is important in both alkaline flooding and alkaline steamflooding. Rock minerals react with and deplete hydroxide ions from the injected alkaline solution. Also, new mineral formation, permeability changes, and scale formations have been observed. Therefore, understanding interactions of alkaline solutions with rock minerals is important for proper reservoir screening and design and operation of floods involving alkali. A detailed study of kaolinite/alkali reaction kinetics and quartz/alkali equilibrium is presented. A small amount of immediate alkalinity loss by ion exchange was observed. At high temperature, the irreversible alkali consumption by kaolinite was very fast for about 100 hours and slow later. Neither hydroxide consumption nor production of dissolved Si or Al followed simple kinetic models of a single irreversible reaction. The kaolinite/alkali reaction (at 180F [82C] and/or 5% NaOH) proceeded incongruently, forming new minerals and consuming proceeded incongruently, forming new minerals and consuming dissolved Si along with hydroxide ions. Therefore, after long-term reaction, the Si concentration dropped to a negligible value even if the initial alkali contained high Si concentration. Under specific conditions of slow reaction, however-e.g., 120F [49C] and 1% NaOH- the alkali consumption could be described by a single first-order reaction. Kaolinite consumed less alkali than montmorillonite at 120F [49C] and 1% NaOH. On the other hand, alkalies equilibrated with quartz contained large amounts of dissolved silica. The equilibrium in the quartz/alkali system was characterized by a (SiO2)/(Na2O) ratio of about 2.0 or by the relationship between concentrations of dissolved silica and hydroxide. The equilibrium results may be applicable to quartzitic sandstones containing insignificant amounts of kaolinite. The complicated nature of solution chemistry involving Si, Al, Na, and other species and the large number of possible minerals render their mathematical modeling extremely difficult.
Present Models for Alkali Loss.
Most of the recent laboratory Present Models for Alkali Loss. Most of the recent laboratory studies and field projects in alkaline flooding used higher alkali concentrations than those used in many earlier projects; the alkali concentration was 1 % or higher in most recent studies. High alkali concentration, combined with high temperature, causes rapid mineral/alkali reactions and fast reduction in hydroxide ion concentration. Hydroxide loss by Na + /H + exchange has been thought to be less than the irreversible loss by mineral/alkali reactions. Two approaches have been used in the past to estimate potential alkalinity loss in the reservoir. One used static (bottle) or flow (core-flood) experiments to obtain a single alkalinity-loss value ieq/100 coring rock. The other, the model approach, used a single irreversible first-order reaction whose rate constant was found from flow tests. The alkali consumption can be underestimated or overestimated in the first approach, depending on the method by which the data are extrapolated for design purposes. For example, if the slow, long-term consumption is ignored, significant underprediction of loss is possible. The second approach should be adequate if the reaction can indeed be modeled by a single irreversible reaction. It appears that rock/alkali reactions are not amenable to modeling by a single irreversible reaction. Experiments with two reservoir sands revealed that the data in one case followed first-order reaction for 4 days but failed to do so in the other case even over such a short time. Another type of mathematical model that allows dissolution/precipitation phenomena is not adequate because of the local equilibrium assumption.
Rock Minerals. Important rock-forming minerals include silica, silicates, alumina, aluminosilicates, carbonates, and sulfates. Most petroleum reservoir rocks are either silica- or carbonate-based. Even though pure calcite does not consume much alkali, carbonate reservoir rocks consume alkali to prohibitive levels because of the reaction of alkali with accompanying gypsum and anhydrite impurities. The silica-based sands and sandstones are composed of quartz, layered aluminosilicates (montmorillonite, kaolinite, illite, chlorite, and mixed-layer minerals), and nonlayered aluminosilicates (feldspars, zeolites, etc.). Quartz is the most abundant fraction by weight and is present largely in the form of sand grains with a fraction present as fines. The clay minerals are crystalline hydrous silicates with a layered structure. Generally, all minerals are present in the crystalline form because of the long geologic age involved. However, some formations may have a fraction of silica present in amorphous form because of lower formation temperatures. XRD is used to obtain the mineral analysis of rocks by weight percent. Perhaps the most important factor of interest is the surface area of individual minerals exposed to the alkaline solution in the pores of formation rock. This information is usually unavailable but may be crudely estimated by scanning-electron-microscope (SEM) identification of clay positions in the formation samples. positions in the formation samples. Montmorillonites (smectite), illite, and kaolinite clays were commonly found in the formation samples. Kaolinite was generally the predominant clay detected, regardless of depth of burial, and predominant clay detected, regardless of depth of burial, and was usually concentrated in isolated spots.
Mineral/Alkali Reactions.
Bunge and Radke reported dissolved silicon concentration from batch dissolution experiments with some minerals using 0.1 N NaOH at 158F [70C], 200 mL alkali/g mineral, and up to 100 hours of reaction. They concluded that the dis-solution of silica is much faster than other minerals and suggested first-order irreversible kinetics. Because of the large liquid/solid ratio and short times involved, however, extrapolation of data to reservoir conditions cannot be justified.
p. 312