The influence of mass transfer on design criteria in liquid extraction (1977)
Type of ContentTheses / Dissertations
Thesis DisciplineChemical Engineering
Degree NameDoctor of Philosophy
PublisherUniversity of Canterbury
AuthorsEspie, A.A. (Anthony Andrew)show all
An investigation into dispersed phase holdup and the droplet size distribution in a counter-current flow packed column, and in particular, into the effects of mass transfer, has been performed using the toluene - acetone - water system in a 15cm ID column with a packed height of 1.40m of 1.6cm OD ceramic Raschig rings.
The holdup data, obtained as a function of flowrates, was qualitatively similar to that of other authors. The presence of a third component resulted in small reductions in holdup due to changes in physical properties except when transfer was out of the dispersed phase. Then, substantial coalescence resulted in reductions in holdup of up to 50%. The data were first analysed using the slip velocity model of Gayler et al (1953). This was approximately obeyed with the effects of continuous phase flowrate not being fully correlated. However, the model of droplet motion used to predict the slope of the slip velocity function was inadequate, especially when mass transfer induced coalescence was occurring. A new model of droplet motion was developed based on the integral volumetric flowrate of all droplet sizes.
The droplet size distributions were controlled by breakup criteria except when transfer was out of the dispersed phase. Then, the coalescence occuring meant that the distribution was an equilibrium between breakup and coalescence. A relatively large distribution was formed at the column inlet and breakup of this occured over more than 60% of the packed height. The impaction mechanism of Ramshaw and Thornton (1967) was important in the breakup of the larger droplets. As the droplets size decreased, restriction breakup by instability and force balance mechanisms became more important. The flow structure within the packing strongly influenced the rate of breakup. These factors were confirmed using a computer simulation to reproduce the experimental distributions.