Surfactant Criteria for Successful Carbon Dioxide Foam in Sandstone Reservoirs

Summary Laboratory results demonstrate that surfactant adsorption on sandstones is minimized and foam performance improved by reducing the ethoxylate chain length in alcohol ethoxy sulfonates and blending unethoxylated and ethoxylated sulfonates to optimize desirable properties. The results also show that laboratory adsorption measurements can only be extrapolated to reservoirs by (1) replicating the anaerobic conditions of reservoirs, and (2) differentiating authogenic minerals from drilling mud found in reservoir cores. Introduction If a foam is to be designed to provide mobility control throughout reservoirs with a thousand meters between wells, the properties of that foam and the surfactants used to create the foam must differ substantially from foams and surfactants used to reduce mobility in near-wellbore applications. First, it is absolutely necessary to maintain very low surfactant adsorption and limited mobility control in order to use foam in the reservoir, yet the foam must survive when the capillary pressure is high. Surfactants with very high critical micelle concentrations (CMC) used below their CMC appear to satisfy these criteria.1 Surfactants used below their CMC have low adsorption and limited mobility control because solid/fluid and fluid/fluid interfaces are not completely filled with surfactant molecules. Then, surfactant adsorption is so low that low concentrations of surfactant can propagate faster than high concentrations.1–3 Yet foam stability and mobility control are sufficient to limit gravity override of gas.1 Second, oil is likely to spread4 around carbon dioxide-rich gas bubbles when they are some distance from an injector because light hydrocarbons have not yet been stripped from the oil. This can mean that another type of foam, gas/oil, is possible.5 Finally, surfactant must somehow propagate where water is not mobile. A good example of this is at the top of an oil reservoir where water saturation is low, water mobility is low, and surfactant does not propagate far in the water. Some evidence suggests5 that the portion of a surfactant which dissolves in the oil does propagate where water is immobile and can stabilize gas/oil lamellae. This gas/oil foam observed in microvisual experiments at reservoir conditions is known to reduce gas mobility when water is absent,5 and shown with simulations to be the likely cause of mobility control in some laboratory experiments.5 This paper describes optimization of surfactant structure and composition as well as the laboratory controls to properly test surfactant performance. The desirable properties of these surfactants are (a) rapid propagation, (b) limited mobility control, and (c) mixed wettability versus water wetness. Because the surfactant is used below its CMC, micelles do not stop chromatographic separation of surfactant components. Experimental Details Experiments were conducted at 170°F (75°C) with the high salinity brines described in Table 1. The surfactant types and equpment used in this study have been described in previous papers.1,5 The surfactants, Fig. 1, were generally alcohol ethoxy [?CH2?CH2 O?] glyceryl [?CH2?CHOH?CH2] sulfonates (AEGS X-Y) supplied by Shell Chemical Co. X and Y refer to the hydrophobe size or range and ethoxylate (EO) number. Results for alpha olefin sulfonates (AOS X) and their mixtures with AEGS surfactants were also discussed. Unethoxylated alkyl biphenyl disulfonates (DOWFAX) were also used. Three types of surfactant adsorption and propagation experiments were conducted. The first, conducted in Berea cores cleaned with a 3 PV 0.1% sodium dithionite rinse, were used to characterize components which propagated with least adsorption. The second group conducted in similarly cleaned Berea cores or slim tubes packed with disagregated reservoir cores were used to optimize surfactant design. The third group, conducted in tubes packed with reservoir sand, were used to demonstrate that surfactants would propagate under reservoir conditions. Some of these were conducted at anaerobic conditions. A limited number of mobility control experiments in slim tubes packed with 2% illite-sand mixture2,5 were also conducted to demonstrate the effectiveness of the surfactants. Surfactant Characterization. The propagating surfactant was initially characterized with 1H nuclear magnetic resonance (NMR), isotachophoresis, and fast atom bombardment mass spectroscopy (FABS) after effluent from a coreflood had been de-salted using the following procedure: One PV of solution from a coreflood was evaporated to dryness and the residue digested in hot methanol to extract the surfactant. The solution was cooled, filtered, and evaporated to dryness. The residue was dissolved in 50/50 H2O then salted out with Na2 CO3 to produce aqueous and isopropanol layers. Water was added to the isopropanol layer and the salting out procedure repeated a second time. The final isopropanol layer was evaporated to dryness to give a de-salted surfactant-rich residue. 1H NMR was conducted at 360 MHz after the surfactant residue had been dissolved in D2O Isotachophoresis was conducted by injecting the surfactant residue into a capillary column in an electric field. Species of different charge are separated and detected by changes of potential gradient at zone boundaries. In FABS, a sample is bombarded by low-energy-charged atoms, which transfer their charge without breaking up the original molecule. The mass of charged species can be determined to four decimal places and used to confirm the identity of a species. Surfactant Adsorption. Surfactant adsorption was measured by injecting a surfactant solution into Berea cores, or 5/8 in.×12 tubes filled with reservoir sand at Sorw. The brines contained sodium dithionite in most of the experiments. The amount of surfactant adsorbed was determined with a methylene blue dye two-phase extraction technique.6 Mineralogy. The mineralogy of reservoir core material was determined by x-ray diffraction and thin-section point counting. The volumes of foreign mud and fines determined from thin-section analysis was subtracted from the total minerals to determine a more realistic reservoir core description.