Visualizing the Effect of Light Oil on CO2 Foams

Summary This paper describes a series of high-pressure, CO2-tertiary-foam, microvisual experiments. The results suggest how oil affects CO2 foam. Foam-stability, emulsion-stability, and spreading-coefficient calculations showhow vaporization and condensation of hydrocarbons in the CO2 process change oilcomposition, spreading behavior of oil at the water/gas interface, and mobilitycontrol. This analysis suggests that water is the spreading phase near theinjector but that oil is the spreading phase elsewhere. Spreading oil benefitsCO2 foam because mass transfer of miscibility-developing components into CO2 ismaintained, while condensation of light hydrocarbons into oil reduces mobilitycontrol enough for reservoir injectivity to be reasonable. A high concentrationof light hydrocarbons in the oil, which causes oil to spread, and oil wetnessappear to be the main reasons that oil impairs foam. Introduction CO2 can effectively recover the oil it contacts, but its displacementefficiency can be reduced because its viscosity and density are lower thanthose of the fluids it displaces. As a result, sweep efficiency declinesbecause of fingering, channeling, and gravity override. Bernard et al.suggested that foam be used to reduce CO2 mobility, thus increasing sweepefficiency. Two useful properties of foam are that a low surfactantconcentration (less than 1,000 ppm) reduces the gas mobility and that foam is anon-Newtonian fluid whose flow resistance decreases with increased velocitynear injectors. Thus, CO2 mobility can be reduced with little added cost, whilethe injectivity constraint thought to result from increasing CO2 apparentviscosity is lifted. A problem that limits foam application is the deleteriouseffect of oil on gas mobility. Some researchers claim that foam fails becauseoil spreads and destroys the foam or because part of the surfactant ispartitioned into the oil. Another problem is that oil recovery can decrease by25 to 50% in low-mobility foams. Because CO2 is not completely miscible with areservoir crude oil, the only region where virtually all the oil is displacedis near the injectors. In a reservoir process, a fraction of an HCPV of CO2 isused. If a foam is to be used as anything but a blocking agent near injectors, the oil's effect must be controlled. To understand how oil affects CO2 foam, weobserved CO2-foam behavior in the presence of reservoir oils in a high-pressuremicrovisual apparatus. Experimental Details Fig. 1 shows the design of the high-pressure visual cell for mountingetched-glass flow cells. The high-pressure cell was designed to achieve highresolution by minimizing refractive index differences. To do this, dimensionswere reduced by using chemically hardened glass windows, by isolating the modelfrom the glycerol overburden with O-rings at the ends of the hollow inlet andoutlet bolts, and by sealing the windows against O-rings with overburden fluid. Microvisual cells were made by silk screening a 0.4 × 3-in. [1 × 7.6-cm]asphalt pattern onto cleaned mirror glass, coating the rest of the plate withasphalt, etching in 25 % hydrofluoric acid, and vigorously scrubbing withsolvents and water. The etched plate was fused to an unetched plate thatcontained the inlet and outlet and was annealed. Silk screening was used todeposit the resistant coating because it is inherently irregular. The modelcould easily be recoated to add more heterogeneity. The etched channels wereabout 100 mu m deep, while the observable PV was about 0.01 cm3. When the inletchannel was filled with glass beads, the total PV was 0.035 cm3. Fig. 2 showsthe complete system. The cell was illuminated with either ultraviolet orvisible fight transmitted obliquely through the model. A stereo microscopeequipped with two phototubes--for video and single-lens reflex (SLR)cameras--was used to view the model. Fluids were injected from sight cells. Theinjection pumps were geared for injection rates between 0.005 and 5 cm3/h. Thebackpressure of the system was maintained by controlling the withdrawal pumprate. Produced volume was determined by measuring the position of thewithdrawal pump with a linear transducer. The produced volume and pressure dropacross the cell were recorded, and the mobilities were calculated. Foamstructure was determined from tracings of photographs digitized with aQuantamet Image Analyzer. Table 1 lists the details of eighteen experimentsconducted at ambient, U.S. gulf coast, and west Texas conditions. In thehigh-pressure experiments, the injection rate was 0.0006 to 0.0012 in.3/hr[0.01 to 0.02 cm3/h], roughly 6 to 12 ft/D [1.8 to 3.7 m/d].