46058-AC9
Non-Equilibrium Nanoblends for Petrochemical Separations
Petroleum refining is the nation's most energy intensive industry and distillations account for 35-40 percent of the total energy consumed. Correspondingly, alternative membrane separations offer potentially enormous energy savings. When robust membranes are developed and deployed in place of distillation, tremendous economic and environmental benefits will occur. Unfortunately, much of membrane science is presently done on a semi-empirical basis; only heuristic guidelines based on rudimentary structure-property relationships are available in the literature. In this project, theories are being developed which are capable of capturing coupling effects, whereby the presence of one component profoundly affects the permeation of another. Such models predict that the nature of specific chemical interactions dictate both solubility and diffusivity selectivity during the separation. The objective of this project is to develop models that link the underlying molecular scale interactions to these recently developed continuum scale permeation models. Both benzene-cyclohexane and butanol-water systems are used for confirming the modeling work experimentally. As part of a comprehensive program of modeling the thermodynamics of the butanol-water system , the Flory free energy of mixing was applied to the butanol-water system. 1-Butanol is of increasing interest as a second generation biofuel due to its availability via fermentation. A two-parameter temperature- and concentration-dependent binary interaction parameter, ?12 = a(T) + b(T)?2, was regressed to the experimental phase envelope. Using the data to fit the bimodal curve in this manner allows calculation of the spinodal and critical points in these systems for the first time. Results were generated for both aqueous solutions of 1-butanol (n-butanol) and 2-(sec-)butanol. It is also proven that the derived chemical potentials obey the Gibbs–Duhem constraint for thermodynamic consistency. A theoretical treatment of multicomponent transport of solutes through crosslinked polymer networks is being developed based upon the principles of non-equilibrium thermodynamics. A multicomponent Flory-Rehner free-energy function provides expressions for chemical potentials and gradients of these potentials serve as the driving forces for permeation. Four different mathematical models (I-IV) have been developed and examined. These four models are based on two distinct forms of the free-energy and two separate assumptions regarding the form of the Onsager coefficients. Models I and II assume the Onsager cross terms are equal to zero; the first uses constant interaction parameters, the second uses concentration-dependent interaction parameters. Models III and IV are also based on constant and non-constant interaction parameters; however, both diagonal and non-diagonal Onsager coefficients are now included. Case study calculations demonstrate that the Onsager cross terms cannot be neglected without compromising the physical nature of the mathematical description. These models are capable of capturing a wide variety of novel permeation phenomena involving the effects of permeant-permeant interaction parameters, permeant-network interaction parameters, permeant molar volume, and the network crosslinking density. The present modeling approach provides a rigorous explanation for differences between ideal selectivities based on pure component permeabilities and observed selectivities resulting when mixed feeds are used. The model based on concentration dependent interaction parameters that includes the Onsager cross terms is found to provide the most comprehensive description of coupled transport in polymer networks.
