This investigation was undertaken to show how the efficiency of solute transfer between two immiscible liquids is influenced by the type, size, and submergence and rotational speed of the impeller and the degree of baffling, residence time, and phase ratio. The system water-kerosene-n-butylamine solute was agitated in a continuous-flow 14 3/4-in. -diam. vessel using propellers, spiral turbines, and flat-blade turbines from 4 to 10 in. in diameter. The feed streams were introduced at the wall at the bottom of the vessel and left at the top of the vessel.
On the basis of the power required for a given level of stage efficiency, the best agitator design is any impeller that has a diameter about 40% of the vessel diameter and is centered in the unbaffled vessel. Without baffles the impeller type and depth of submergence have little effect on performance. The power increase required by the addition of baffles is small at the highest stage efficiencies, especially with the larger radial-flow impellers, but may be severalfold at efficiency levels of 70 to 90%. This presumably results from the lowered mass transfer driving force caused by the increase in end-to-end mixing due to wall baffles. Whether baffles have such a large adverse effect if the feed is introduced into the impeller rather than at the vessel wall is not fully established, but there is some evidence that baffling is also undesirable in this case. For baffled operation, impeller type and submergence do affect performance.
A correlation of stage efficiency is presented in terms of Reynolds number and power number. In addition, energy input per volume of total flow is correlated in terms of residence time and stage efficiency for one size of baffled propeller.
Spot samples taken from the vessel showed large, random-concentration fluctuations out to 20 residence times and perhaps indefinitely. Reproducible results were obtained by taking time-averaged samples.