Impact of Surfactant Mixtures on Microemulsion Phase Behavior

Abstract

Abstract Surfactant feedstock for surfactant-polymer (SP) flooding field applications is commonly delivered as a blend of surfactant mixtures with a distribution of molecular weight (MW), rather than as a pure component. Moreover, two or more synthetic commercial surfactants with their own MW distributions are often injected simultaneously. In order to design an effective SP flooding process it is important to predict phase behavior of complex surfactant mixtures with changing formulation variables, such as temperature, pressure, oil EACN, brine salinity, and surfactant composition. The main challenges to model phase behavior of surfactant mixtures include distinct optimum conditions for each type of surfactant, limited information about the distribution of the surfactants in the commercial blends, different partitioning of surfactants in the excess phases and microemulsion phases, and dynamically altered surfactant composition due to selective adsorption, degradation, or differing partition coefficients. In this paper, we present new mixing rules for our recently developed equation of state model to predict phase behavior of surfactant mixtures with changing compositions of two or more synthetic commercial grade surfactants. The model is based on the hydrophilic-lipophilic difference (HLD), net-average curvature (NAC), and mole fraction weighted ratios of surfactant blends (in terms of both multiple synthetic surfactants and their individual distributions). Significant chromatographic separation of surfactant mixture may occur as a result of distinct adsorption isotherms for different surfactant components. These composition changes can have a critical impact on the phase behavior of the mixture and the ultimate oil recovery. Fish plots are constructed for varying water-oil ratios and surfactant compositions that show significant sensitivity to water-oil ratio and surfactant composition. When the injected chemicals do not separate (e.g. there is no selective adsorption, degradation, or selective partitioning), fish plots for a water-oil ratio of 1.0 are symmetric with surfactant concentration because the surfactant blend behaves as a pure surfactant pseudocomponent. For water-oil ratios not equal to 1.0, the surfactant blend can still be modeled as one pseudocomponent, as long as surfactant composition remains fixed, but the fish plot becomes asymmetric. Fish plots, however, are always asymmetric when surfactant composition changes. In such cases, the surfactants must be represented by multiple components with different characteristics and distributions. The results are illustrated using ternary and quaternary diagrams based on measured experimental data in the literature. The new model is included in our in-house compositional simulator PennSim, which shows that simulations of a surfactant flood with different adsorption isotherms can significantly change recovery.