(174b) A Modeling Strategy for Optimal Solvent Composition Selection in the Design of a New Api Process | AIChE

(174b) A Modeling Strategy for Optimal Solvent Composition Selection in the Design of a New Api Process

Authors 

Hsieh, D. - Presenter, Bristol-Myers Squibb Company
Corbett, M. - Presenter, Joule Unlimited
Wei, C. - Presenter, Bristol Myers Squibb
Wang, S. S. - Presenter, Bristol-Myers Squibb Company
Kiang, S. - Presenter, Bristol-Myers Squibb Company


We studied the solvent compositions for the reaction-L/L extraction-distillative crystallization sequence featured in a new API process, to further define the operable parameter space. Three liquid components were considered, one of which was aqueous. The organic components were isobutanol and toluene.

During reaction, two liquid phases were required and impurities generated in the organic rich phase could be continuously extracted into the aqueous rich phase. The amount of aqueous component required to generate two liquid phases was dependent upon the weight ratio of solvent isobutanol to toluene and the miscibility of both solvents with the aqueous component. The boundaries defining these relationships were predicted by Aspen Plus through creation of a ternary phase diagram.

Following the reaction, additional water was introduced to extract the residual impurities in the organic phase. The organic rich phase after extraction included all three components. However, crystallization in toluene was desired. Aspen Plus was used to generate the residue curve map and distillation boundary for this phase which featured three binary azeotropes and one ternary azeotrope.

Each of the three unit operations required a different solvent composition space (solvent design space) in the ternary phase diagram in order to achieve optimal operation. A general modeling approach was utilized to locate a ?sweet spot? in the ternary phase diagram from which the transitions between the solvent compositions as the process progressed could be greatly facilitated.