A molecular-based group contribution equation of state for the description of fluid phase behaviour and thermodynamic derivative properties of mixtures (SAFT-γ Mie)

摘要

An accurate knowledge of the thermophysical properties and phase behaviour ofudfluid mixtures is essential for the reliable design of products and processes across a wideudrange of chemical engineering applications, varying from the processing of petroleumudfluids to the manufacturing of pharmaceuticals. Thermodynamic tools and, in theudcontext of this work, group contribution (GC) methods are predictive approachesudthat are expected to play an important role in meeting these industrial needs. Theudprincipal focus of the work presented in this thesis is the development of a novel GCudmethod based on the statistical associating fluid theory (SAFT): the SAFT-γ Mieudapproach. The method is developed based on a detailed molecular model and a realisticudintermolecular potential, the Mie potential with variable attractive and repulsiveudranges, for the description of interactions at a molecular level. Over the past decade,udan increasing research effort has been devoted to developing formalisms that coupleudthe accuracy of the SAFT equation of state (EoS) with the predictive capabilities ofudgroup contribution approaches. In the development of such methods one aims to overcomeudthe limitations inherent to GC approaches based on activity coefficient models,udsuch as in the well-established universal quasi-chemical functional group activity coefficientud(UNIFAC) approach. A more recent landmark has been the development ofudheteronuclear methods within SAFT. The SAFT-γ EoS based on the square-well (SW)udpotential has been shown to describe accurately the phase behaviour of a wide varietyudof fluids. In the work presented in this thesis, SAFT-γ SW is applied to the study ofudthe fluid phase behaviour of aqueous solutions of hydrocarbons. These mixtures are ofudhigh industrial relevance, and the accurate representation of their highly non-ideal natureudis very challenging from a theoretical perspective. The SAFT-γ method is shownudto perform comparatively well in predicting the behaviour of the systems examined.udNonetheless, some challenges are identified, such as the description of thermodynamicudderivative properties and the description of near-critical fluid phase behaviour, whereudthe performance of the method is shown to be less accurate. These challenges partiallyudarise from the simplistic intermolecular square-well potential employed within SAFT-γudSW, which allows for a rigorous theoretical development, but fails to reproduce accuratelyudfiner aspects of the thermophysical behaviour of fluids, such as second-orderudderivative thermodynamic properties.udThese challenges are tackled here with the development of the SAFT-γ Mie GCudapproach, based on the versatile Mie intermolecular potential and a third-order treatmentudof the thermodynamics of the monomer segments. The SAFT-γ Mie methodudis applied to the study of the properties of two chemical families, n-alkanes and 2-udketones, and it is shown that a significant improvement over existing SAFT-basedudgroup contribution approaches can be achieved in the description of the pure componentudphase behaviour of the compounds studied. Moreover, the application of a realistic intermolecular potential is shown to allow for an excellent description ofudsecond-order derivative thermodynamic properties, and the accurate treatment of theudintersegment interactions is shown to improve the performance of the method in theuddescription of the near-critical fluid phase behaviour. The predictive capability of theudmethod is demonstrated in the description of mixture fluid phase behaviour and excessudthermodynamic properties in a predictive manner. Given the promising performanceudof the SAFT-γ Mie EoS, the method is applied to the case study of the solubilityudof two active pharmaceutical ingredients in organic solvents. The method is shownudto satisfactorily predict the solubilities of the mixtures considered, based on limitedudexperimental data for simple systems. Given the complexity of the mixtures studied,udthe performance of the SAFT-γ Mie is considered very encouraging and shows thatudthere is great potential in the application of the method to this challenging field.