Measurements, Modeling, and Database Development for the Application of Alternative Solvents

T.J. Bruno and A.F. Lagalante

Objective: To develop and test predictive models for solvation of compounds in alternative solvents at supercritical, near-critical, and subcritical conditions using a combination of both physical and chemical variables as input into an empirical multivariate statistical model.

Problem: The most important thermophysical parameter required to assess the feasibility of an extraction process is the solute-solvent phase equilibrium. Serious limitations exist in equation-of-state approaches that use only physical properties of the solute and solvent to model phase equilibrium.

Approach: Safe replacements for conventional solvents are likely to come from fully or partially fluorinated alkanes, ethers, or ketones that possess negligible ozone depletion potential, as well as functionalized glycol ethers and siloxanes. Many of the fluorinated alternative solvents are gases under ambient conditions, and their thermophysical properties offer the promise of both conventional liquid extraction and the tunable solvent strength offered by near-critical and supercritical fluid extraction. In our approach, a given solution process is empirically modeled as the dependent variable in a multivariate statistical analysis. The independent variables to the multivariate statistical model include empirical solute-solvent interactions and additional state-dependent terms. Solute-solvent interactions are quantified using empirical solvatochromic and chromatographic parameters of acidity, basicity, polarizability, and polarity. These parameters represent the dominant chemical interactions in solvent-solute systems and will account for contributions to the nonideal portion of phase equilibrium. Accounting for these interactions will permit higher accuracy than EOS approaches. The statistical model aids in the identification of alternative solvents by making it possible to predict the solubility of industrially relevant compounds.

Results and Future Plans: In recent years, we have designed and constructed numerous spectroscopic, chromatographic, and gravimetric instruments for the measurement of solubilities of solutes in sub- and supercritical fluids. Solutes studied have ranged from classes of organometallic compounds to physiologically active natural products. The solvatochromic parameters for the fluorinated ethane solvents have been measured using high-pressure spectroscopic cells. Values are density-dependent over the gas-to-liquid density range and have been used to model R143a/water, R134a/water, and carbon dioxide/water partitioning of organic solutes. Parameters for the glycol ethers, alkanolamines, and siloxanes have been measured for both the pure compounds and aqueous solutions. Soon, we will be developing a fiber-optic solvatochromic sensor to facilitate measurement of solvatochromic parameters. The sensor will allow the rapid measurement of solvent mixtures for determination of mixture parameters. Pure component and mixture parameters will be incorporated into a database that will allow researchers to statistically model an industrially pertinent solvent replacement technology. The database model will suggest suitable alternative solvents and extraction conditions to substitute for a hazardous solvent.

Publications:

Lagalante, A.F. and Bruno, T.J., "Modeling the Water-R143a Partition Coefficients of Organic Solutes Using a Linear Solvation Energy Relationship," J. Phys. Chem. 103, 7319 (1999).

Lagalante, A.F., Hall, R.L. and Bruno, T.J., "Kamlet-Taft Solvatochromic Parameters for the Fluorinated Ethane Solvents," J. Phys. Chem. B 102, 660 (1998).