EFFECT OF FREE CONVECTION ON THERMODIFFUSION IN A LIQUID MIXTURE FILLING AN INCLINED RECTANGULAR CAVITY K. R. Kostarev and A. F. Pshenichnikov UDC 536.25 The convective motion which develops in an inclined cavity upon heating from above determines to a significant degree the form of the concentration field produced by thermodiffusion. The interaction of convective and thermodiffusion fluxes at small thermal Grashof numbers Gr causes the appearance of longitudi- nal jumps in concentration. Increase in temperature difference intensifies convection and encourages reduction in concentration gradients. The dominant role of convection for fixed Gr is determined by the angle of inclination of the liquid layer [i, 2]. A significant feature of liquid solutions is their low diffusion coefficient and thus high Schmidt number. This fact does not permit use of results obtained for gas mixtures, and greatly complicates nu- merical simulations. In contrast to [2], the present study will investigate thermodiffusion separation in a cavity with impermeable boundaries. The rectangular cavity considered (Fig. I, T 2 > T i) was a gap between planar brass heat exchangers formed by use of two Plexiglas inserts ~ = 2.0 mm thick. The distance be- the inserts was 30.0 mm. On the endfaces the gap was closed by two plane semitransparent mirrors. The mirrors were set parallel to each other using alignment screws, forming the working space of a two-mirror autocollimation interferometer, which allowed study of concen- tration fields [3, 4]. The distance between mirrors was 30.3 mm. Interferograms obtained under isothermal conditions depicted lines of equal concentration. The transition from one interference band to the next corresponded to a change in concentration of the working solution (solution of sodium sulfate in water) of 3.4.10-3%. The apparatus was equipped with a device which allowed changing its angle to the horizontal over the range of -15 to 60 ~ , the axis of rotation being perpendicular to the plane of the mirrors. The amount of inclination was determined with an optical quadrant with scale divisions of 30". The temperature of the heat exchangers was controlled by two jet-type ultrathermostats, provided with additional temperature stabilization systems. Temperature change was measured by a copper-Constantin differential thermocouple, connected to a type Shch-300 digital volt- meter. A mercury thermometer with scale divisions of 0.1~ was used to measure the tempera- ture of the surrounding medium. The working liquids used were 6.62% and 15.7% aqueous solutions of sodium sulfate, which differ from solutions of other salts in their high Soret coefficient and low diffusion coefficient [5]. This latter fact permits their use in study of concentration fields by the following technique. At the beginning of the experiment the model is oriented at some angle to the horizontal. The heat exchangers are connected to the thermostats. Under the action of the temperature difference convection develops in the cavity, and after the pass- age of some time (about 1.5 h) a steady-state concentration field is established. The time required for formation of a steady state temperature field did not exceed 6 min. Before measuring concentration changes both heat exchangers were connected to the cold thermostat, thermal convection was halted, and the slow process of equalization of concentra- tion perturbations began, mainly because of diffusion. Because the temperature perturbation relaxation time is two-three orders of magnitude less than the concentration perturbation relaxation time, the equalization takes place under isothermal conditions (temperature dif- ferential did[ not exceed 0.05~ Interferograms obtained under these conditions were used to track the changes in the concentration field, and its parameters, corresponding to the initial steady-state regime, were calculated. To do this the experimentally obtained depend- Perm'. Translated from Zhurnal Prikladnoi Mekhaniki i Tekhnicheskoi Fiziki, No. 5, pp. 73-76, September-October, 1986. Original article submitted August 8, 1985. 0021-8944/86/2705-0695512.50 9 1987 Plenum Publishing Corporation 695