Mario Blanco, William A. Goddard III

Materials and Process Simulation Center

MSC 99 Research Conference

Caltech, March 19, 1999




There is a tremendous potential for dramatic reductions in pollution production through optimization of chemical manufacturing process design. Chemical equilibrium constants (Kc), reaction rates (k), and activity coefficients (gi) are the three most fundamental properties needed to calculate optimal solutions to common engineering problems such as mixing and separations (formulation, distillations, solvent extractions), and catalysis (hydrodesulfurization (HDS), hydrodearomatization (HDA), hydrodenitrogenation (HDN)) to name a few. Chemical engineering databases are not capable to provide all the required information, as the number of parameters increases exponentially with the number of mixture components.


The process of experimentally determining activity coefficients is an on going and difficult process into which millions of dollars of research is spent each year. Understandably, many semi-empirical methods have been proposed to predict these values, of which, UNIFAC, the UNIversal Functional group Activity Coefficient method, has arguably been the most successful. Group additivity estimation methods, such as UNIFAC, still require costly experimentation to refine the parameters needed to make useful predictions. Often group additivity methods lead to ambiguous results depending on how chemical compounds are represented in terms of individual chemical groups.


We outline how it is now possible to obtain all three properties (Kc, k, and gi) from fundamental theory, without the use of experimental data or the use of group additivity methods. We call this de novo Chemical Engineering. We will focus on the level of theory required as well as on the most common pitfalls encountered by the expert.