Turbulent mixing and the selectivity of fast chemical reactions

Abstract

Assuming that complete information is available about (a) the chemical reactions (mechanism, stoichiometry, thermodynamics, and kinetics) and (b) the turbulent flow field in a reactor (mean velocities and turbulence parameters), we would like to calculate directly the time evolution of the temperature and the concentrations of all species present. No general method is yet available to satisfy this goal. Starting from the phenomena of reaction, molecular diffusion, and laminar deformation in laminated structures, which are formed by the action of vorticity, mixing in liquids at sub-Kolmogorov scales can be quantitatively described. This approach requires the compositions of the fluids incorporated into the vortices to be known. Thus, for example, when a reagent solution (B) is slowly added to a huge volume of a second reagent solution (A) in a stirred tank reactor, fresh B mixes with the average A composition. (Slowly means that the addition time is much greater than the time needed for turbulent and molecular transport to homogenize the tank contents.) Progress of the reactions is then determined essentially by the interaction between the fine-scale mixing (near the Batchelor concentration scale) and the chemical kinetics. A convenient way to evaluate this modeling compares predicted and measured product distributions (or selectivities) for fast multiple reactions. (Up to now, this has been done for dilute solutions, which exhibited concentration, rather than also temperature segregation.) Advantages compared to working with single reactions can include enhanced sensitivity, inclusion of a wider range of flow fields than turbulent pipe flow, e.g., well-stirred systems, and practical relevance of controlling product distributions in the manufacture of chemicals. The effects on the product distributions of fast reactions of operating parameters such as stoichiometric and volumetric feed ratios, reagent concentrations, feed location, impeller speed, size and shape, viscosity, and type of reaction can be well described. Present efforts concentrate on relaxing the assumption of slow feed addition and of describing reactors where not only fine-scale, but also coarse-scale inhomogeneities occur. A simplified model for the engulfment frequency between fluids having different compositions is being evaluated. Some progress has been made, but the way in which fluid elements mix with each other is still not fully known.