Luke D. Simoni

Extraction of alcohols from water with 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide

Ethanol production in the U. S. has increased 36% between 2006 and 2007 (J. M. Urbanchuk, Contribution of the Ethanol Industry to the Economy of the United States, LECG, LLC, Renewable Fuels Association, 2008) in response to a growing demand for its use as a commercial transportation fuel. 1-Butanol also shows potential as a liquid fuel but both alcohols require high energy consumption in separating them from water. 1-Butanol, in particular, is considered an excellent intermediate for making other chemical compounds from renewable resources, as well as being widely used as a solvent in the pharmaceutical industry. These alcohols can be synthesized from bio-feedstocks by fermentation, which results in low concentrations of the alcohol in water. To separate alcohol from water, conventional distillation is used, which is energetically intensive. The goal of this study is to show that, using an ionic liquid, extraction of the alcohol from water is possible. Through the development of ternary diagrams, separation coefficients are determined. The systems studied are 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide/ethanol/water, which exhibits Type 1 liquid–liquid equilibrium (LLE) behavior, and 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide/1-butanol/water, which exhibits Type 2 LLE behavior. Based on the phase diagrams, this ionic liquid (1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide) can easily separate 1-butanol from water. It can also separate ethanol from water, but only when unreasonably high solvent/feed ratios are used. In addition, we use four excess Gibbs free energy (gE) models (NRTL, eNRTL, UNIQUAC and UNIFAC), with parameters estimated solely using binary data and/or pure component properties, to predict the behavior of the ternary LLE systems. None of the models adequately predicts the Type 1 system, but both UNIQUAC and eNRTL aptly predict the Type 2 system.

Extraction of biofuels and biofeedstocks from aqueous solutions using ionic liquids

Abstract The production from biomass of chemicals and fuels by fermentation, biocatalysis, and related techniques implies energy-intensive separations of organics from relatively dilute aqueous solutions, and may require use of hazardous materials as entrainers to break azeotropes. We consider the design feasibility of using ionic liquids as solvents in liquid–liquid extractions for separating organic compounds from dilute aqueous solutions. As an example, we focus on the extraction of 1-butanol from a dilute aqueous solution. We have recently shown ( Chapeaux et al., 2008 ) that 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide shows significant promise as a solvent for extracting 1-butanol from water. We will consider here two additional ionic liquids, 1-(6-hydroxyhexyl)-3-methylimidazolium bis-(trifluoromethylsulfonyl)imide and 1-hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate, as extraction solvents for 1-butanol. Preliminary design feasibility calculations will be used to compare the three ionic liquid extraction solvents considered. The ability to predict the observed ternary liquid–liquid equilibrium behavior using selected excess Gibbs energy models, with parameters estimated solely using binary data and pure component properties, will also be explored.

Process design using ionic liquids: Physical property modeling

Abstract Ionic liquids are a relatively new class of materials with properties that make them attractive for a wide variety of engineering applications. For design purposes, it is useful to have a relatively simple model (i.e., excess Gibbs energy model or equation-of-state model) capable of describing the physical properties and equilibrium behavior of ILs and IL solutions. We consider here the performance of two selected models, NRTL applied to the modeling of liquid-liquid equilibrium and an electrolyte equation-of-state applied to the modeling of aqueous mean ionic activity coefficients. In each case we focus on issues in parameter estimation, and use an approach based on interval mathematics to solve the parameter estimation problem globally. Sample results are presented and suggest that the models considered here may be useful for correlation of data, but may not be well suited for prediction.