Effect of solvent on controllability in extractive distillation

There have been numerous books and papers dealing with the effect of solvent selection on the steady-state design of extractive distillation systems. Extractive distillation is widely used to separate homogeneous binary azeotropes by adding a higher-boiling solvent that attracts one of the components. These systems typically feature an extractive column and a solvent-recovery column. The two products are removed in the distillate streams from the two columns. The solvent is circulated from the base of the solvent recovery back to near the top of the extractive column. A recent paper by Kossack et al. (Chem. Eng. Res. Dev., in press) discusses in detail the steady-state economic design issues and presents several examples that use a variety of solvents for the separation of acetone and methanol. Some solvents drive the acetone overhead in the extractive column, while others drive the methanol overhead in the extractive column. This paper extends this work to compare the dynamic performance of the acetone/methanol system with different solvents. The numerical cases are based on those presented by Kossack et al. Three solvents are explored: water, dimethyl sulfur oxide, and chlorobenzene. The first and second solvents drive the acetone overhead in the extractive column. The last solvent drives the methanol overhead in the extractive column. Control structures are developed for each system. Dynamic simulation results show that all systems are controllable, but product quality variability is poorest when the methanol is driven overhead in the extractive column. One explanation for this difference is that methanol is higher boiling than acetone and preferentially wants to go out the bottom.