Date of Award
Doctor of Philosophy (PhD)
Dr. A. Benedek
Although adsorption of organic contaminants from the aqueous phase onto activated carbon is a widely used process, mechanistic models capable of describing the performance of adsorption beds are not presently available. Such models, if available, would allow an improved understanding of the behaviour of adsorption beds and could reduce the number of pilot plant studies which are presently required for system design. To develop a mechanistic model of activated carbon adsorption a two part study was undertaken.
In the first part, existing adsorption kinetic models were investigated and were found to be incapable of describing the adsorptive behaviour. Based on the structural properties of activated carbon, a branched pore kinetic model is presented which separates the carbon particle into two regions, a rapidly diffusing region which is associated with pores which are significantly larger than the adsorbing molecules, and a slowly diffusing region which is associated with pores which are of a comparable size to the diffusing molecules and in which diffusion is retarded. Equilibrium and batch kinetic experiments were conducted with pure phenolic compounds as adsorbates, and the model parameters were calculated from the data. The analysis shows that the slow diffusion region is important, and that some previously inconsistent results reported in the literature, such as premature breakthrough and tailing of the breakthrough curves, are due to the neglect of this factor. When used in a model of an adsorption column, the kinetic parameters calculated from the batch experiments enabled the performance of experimental adsorption columns to be predicted with excellent accuracy over extended periods. The early breakthrough and characteristics tailing noted above are shown to be a result of uptake in the slow diffusion region. The model as developed originally assumed a surface diffusion mechanism was responsible for transport in the rapidly diffusing region and as a result, solution of the model required extensive computation. The use of the quadratic driving force assumption was investigated to replace the surface diffusion mechanism and its use is shown to significantly reduce the solution complexity while giving an equally good interpretation of the data.
In the second part of the study the adsorptive behaviour of two feedstreams of environmental importance was studied. As in the first part, equilibrium and batch kinetic experiments were conducted and parameters for the quadratic driving force branched pore model were obtained by comparing the model to the data. Adsorption column experiments were conducted over extended periods and when the batch determined parameters are included in the adsorption bed model, the bulk of the breakthrough curve is shown to be well predicted. Because of the complexity of the feedstreams treated, the concentration can only be quantified in terms of a non-specific parameter such as Total Organic Carbon (TOC) and all the kinetic constants must be evaluated in terms of this parameter.
As with the pure solutes in the first part of the study, the slow diffusion region is very important in the absorptive uptake of these complex organic contaminants. Removal by biological mechanisms is often suggested when adsorption columns continue to exhibit an uptake capacity after prolonged periods of operation, and the problems in differentiating between very slow adsorptive uptake and biological removal are discussed in detail. The present analysis shows that in many cases this removal may be due to continued slow adsorption rather than biological oxidation of the organics. The slow diffusion also has important consequences for the design of activated carbon contactors as contact times far in excess of current practice are indicated if maximum use is to be made of the adsorptive capacity.
Peel, Russell G., "The Roles of Slow Adsorption Kinetics and Bioactivity in Modelling of Activated Carbon Adsorbers" (1979). Open Access Dissertations and Theses. Paper 647.