Tuesday, November 6, 2007 - 3:45 PM
304c

Transport Limitations in Thermal Diffusion

Nicholas Cox1, Pawel Drapala2, and Bruce A. Finlayson1. (1) Chemical Engineering, University of Washington, Dept. Chemical Engr. Box 351750, University of Washington, Seattle, WA 98195, (2) Chemical Engineering, Illinois Institute of Technology, Chicago, IL

Numerical simulations are made of thermal flow field flow fractionation to compare with experimental data in the literature. In this technique, molecular diffusion and thermal diffusion compete to create a concentration gradient in an otherwise uniform concentration field. In any flow device there are entrance regions in which the velocity and temperature profile are developing. Thus, interpretation of data is clouded by the fact that the temperature gradient is not established over the total region of flow. The magnitude of such errors is determined for flow devices described in the literature.

In a thermal diffusion cell, used to measure the Soret coefficient, the ratio of the thermal diffusion coefficient to the molecular diffusion coefficient, the profile is developed over some time. A macroscopic model provides guidance about the time to reach steady state.

The same phenomenon is relevant in the experiments by Braun and Libchaber (Phy. Rev. Letters, vol. 89, 188103 (2002), “Trapping of DNA by Thermophoretic Depletion and Convection.” In this experiment, a thin cylinder of fluid containing DNA is heated by a laser along the centerline. Convection is created through a gravitational force, and a temperature gradient is established. This temperature gradient then provides a driving force for thermal diffusion, which causes the DNA to increase in concentration near the center and bottom of the cylinder. Simulations show the same behavior as seen experimentally.