No abstract

No abstract

of phase equilibria is fundamental to the understanding of heterogeneous systems and to their use in separation processes and other chemical engineering operations. Phase equilibria in the critical region-the region a t elevated pressures above the critical temperature of one of the compounds-are characterized by large nonidealities in the gas phase as well as possible liquid phase nonidealities. These nonidealities are manifested in interesting but generally unpredictable phenomena such as the solution of relatively large amounts of nonvolatile liquids and solids in supercritical gases, the existence of liquid phases consisting almost entirely of a supercritical component, phase inversions, gas-gas immiscibilities, etc. Unique separation processes might utilize the large changes in solubilities that occur over small temperature and pressure ranges in the critical region.The present investigation is part of a continuing program a t Princeton aimed a t the understanding and generalization of phase equilibria just above the critical temperature of the gaseous component. Todd and Elgin (16) presented results of the original investigation of organic liquids with supercritical ethylene (T,9.5" C.) which established the three types of pressure-composition isotherms that have also been found in all subsequent binary systems. This work was extended by Gottschlich ( 7 ) , and ternary systems with ethylene were studied by Elgin and Weinstock ( 5 )

The rational design of a reaction system to produce a desired polymer is more feasible today by virtue of mathematical tools which permit one to predict product distribution as affected by reactor type and conditions. New analytical tools such as gel permeation chromatography are beginning to be used to check technical predictions and to aid in defining molecular parameters as they affect product properties. The vast majority of work concerns bulk or solution polymerization in isothermal batch or continuous stirred tank reactors. There is a clear need to develop techniques to permit fuller application of reaction engineering to realistic nonisothermal systems, emulsion systems, and systems at high conversion found industrially. A mathematical framework is also needed which will start with carefully planned experimental data and efficiently indicate a polymerization mechanism and statistical estimates of kinetic constants rather than vice-versa./^wing to a recent increase in theoretical treatments of polymerization kinetics, processes, and reactor design, chemical engineers are becoming increasingly active in a field where physical chemists primarily published. This paper reviews the theoretical tools now available and attempts to identify the work needed to make a fundamental approach more useful in practice. The broader field of chemical reactor engineering is itself relatively new, with the first modern text by Brotz (10) and by Walas (56) appearing in 1958 and 1959, respectively. Of the several thousand pages that have appeared since 1958, however, only a few deal with polymerization reactors, hardly sufficient recognition of their commercial importance. The recent activity seems triggered by the avail-'Also staff member,

Fickian and non‐Fickian diffusion into polymer‐solvent systems were simulated on both analog and digital computers. Methods were evaluated to solve the descriptive nonlinear ordinary and partial differential equations. Analog methods published previously have been expanded. It was shown that discrete time solutions are superior to discrete distance solutions for partial differential equations, as hardware requirements are small. Also, accuracy problems on some of the smaller installations can be reduced by reverse time integration. For highly nonlinear equations, analog methods are superior to digital in stability and effort required. Digital solutions can be require extensive preliminary work to define computation parameters. However, the errors introduced by the finite difference equations are small in moderately nonlinear equations and much more accurate solutions are possible. Generalized solutions for polymer‐solvent diffusion with exponential concentration dependence are presented for the Fickian cases. Also, typical solutions are demonstrated for all known types of non‐Fickian diffusion.