Experimental and theoretical investigation of Fickian and thermal diffusion coefficients in hydrocarbon mixtures

Fickian diffusion is the passive movement of molecules due to composition gradient in a mixture, and is quantified by Fickian diffusion coefficients Dm. Thermal diffusion is the selective movement of molecules in the presence of temperature gradient, and is quantified by thermal diffusion coefficients, DT. Multicomponent Fickian and thermal diffusion have proven importance in areas such as isotope separation, combustion, and in various aspects of petroleum engineering, for example wax deposition in pipelines, improved oil recovery, and hydrocarbon reservoir initialization. In this work, we investigate Dm and DT using both experimental and theoretical approaches. We measure DT and Dm for binary and ternary liquid hydrocarbon mixtures. A comparison of the binary and ternary DT reveals a remarkable difference in the thermal diffusion behavior in binary and ternary mixtures. We demonstrate that ternary mixtures may not be considered pseudo-binary mixtures, and cross-Fickian diffusion coefficients may not be neglected. Results for binary mixtures show a significant effect of molecular shape and size on both coefficients. We also show that in binary mixtures, DT and mixture viscosity, both non-equilibrium properties, are closely related. Fickian diffusion coefficients are affected by fluid non-ideality and the multicomponent nature of mixtures. There is currently no single method that accurately predicts the Dm matrix in non-ideal gas and liquid multicomponent mixtures. In this work, we develop a model, motivated by the corresponding states approach, for non-polar multicomponent gas and liquid mixtures. We use the new model to calculate the dependency of D m on composition, pressure and temperature for multicomponent mixtures, using a generalized multicomponent Vignes relation and a description of mixture non-ideality in the framework of irreversible thermodynamics, using the volume-translated Peng-Robinson equation of state. Our model predicts Dm for non-polar mixtures with an absolute average deviation of 12% for 889 data points. We provide a simple expression that clearly shows the behavior of DT at the critical point. We also suggest an improvement to the current predictive model for DT, based on irreversible thermodynamics and a molecular understanding of the thermal diffusion process